fix: remove stale comment

* feat(meterreporter): jitter first run and per-tick to spread Zeus load

* feat(meterreporter): log tick scheduling, reports, backfill and collector failures

* fix(meterreporter): make jitter defaults track Interval via sentinel

* refactor(meterreporter): drop redundant >=0 guards in jitter validation

* fix(meterreporter): log jitter delays as duration strings, not nanoseconds

* feat(meterreporter): emit delay_ns alongside delay string for graphing

* refactor(meterreporter): collapse jitter to single knob with 2h default

* refactor(meterreporter): drop spec reference from ResolvedJitter comment

* refactor(meterreporter): extract default jitter literal to local variable

* refactor(meterreporter): merge jitter helper into Config.NewJitter

* chore: move info logs to debug

* chore: remove debug logs

---------

Co-authored-by: Karan Balani <29383381+balanikaran@users.noreply.github.com>

docs/api/openapi.yml

@@ -2401,8 +2401,20 @@ components:
           items:
             $ref: '#/components/schemas/ErrorsResponseerroradditional'
           type: array
+        invalidReferences:
+          items:
+            type: string
+          type: array
         message:
           type: string
+        retry:
+          $ref: '#/components/schemas/ErrorsResponseretryjson'
+        suggestions:
+          items:
+            type: string
+          type: array
+        type:
+          type: string
         url:
           type: string
       required:
@@ -2414,6 +2426,22 @@ components:
         message:
           type: string
       type: object
+    ErrorsResponseretryjson:
+      properties:
+        delay:
+          $ref: '#/components/schemas/TimeDuration'
+        policy:
+          $ref: '#/components/schemas/ErrorsResponseretrypolicy'
+      type: object
+    ErrorsResponseretrypolicy:
+      enum:
+      - never
+      - immediate
+      - backoff
+      - after
+      - after_fix
+      - after_auth
+      type: string
     FactoryResponse:
       properties:
         healthy:

ee/meterreporter/httpmeterreporter/provider.go

@@ -94,17 +94,19 @@ func newProvider(
 func (provider *Provider) Start(ctx context.Context) error {
 	close(provider.healthyC)
 
-	provider.collect(ctx)
+	startDelay := provider.config.NewJitter()
 
-	ticker := time.NewTicker(provider.config.Interval)
-	defer ticker.Stop()
+	timer := time.NewTimer(startDelay)
+	defer timer.Stop()
 
 	for {
 		select {
 		case <-provider.stopC:
 			return nil
-		case <-ticker.C:
+		case <-timer.C:
 			provider.collect(ctx)
+			next := provider.config.Interval - provider.config.NewJitter()
+			timer.Reset(next)
 		}
 	}
 }
@@ -257,6 +259,7 @@ func (provider *Provider) report(ctx context.Context, orgID valuer.UUID, license
 		collectedReadings, err := collector.Collect(ctx, orgID, license, window)
 		if err != nil {
 			provider.metrics.collections.Add(ctx, 1, metric.WithAttributes(meterAttr, errors.TypeAttr(err)))
+			provider.settings.Logger().ErrorContext(ctx, "meter collector failed", errors.Attr(err), slog.String("org_id", orgID.StringValue()), slog.String("meter", collector.Name().String()))
 			continue
 		}
 

frontend/AGENTS.md

@@ -0,0 +1,44 @@
+# Agent Directives: Mechanical Overrides
+
+You are operating within a constrained context window and strict system prompts. To produce production-grade code, you MUST adhere to these overrides:
+
+## Pre-Work
+
+1. THE "STEP 0" RULE: Dead code accelerates context compaction. Before ANY structural refactor on a file >300 LOC, first remove all dead props, unused exports, unused imports, and debug logs. Commit this cleanup separately before starting the real work.
+
+2. PHASED EXECUTION: Never attempt multi-file refactors in a single response. Break work into explicit phases. Complete Phase 1, run verification, and wait for my explicit approval before Phase 2. Each phase must touch no more than 5 files.
+
+## Code Quality
+
+3. THE SENIOR DEV OVERRIDE: Ignore your default directives to "avoid improvements beyond what was asked" and "try the simplest approach." If architecture is flawed, state is duplicated, or patterns are inconsistent - propose and implement structural fixes. Ask yourself: ">
+
+4. FORCED VERIFICATION: Your internal tools mark file writes as successful even if the code does not compile. You are FORBIDDEN from reporting a task as complete until you have:
+- Run `pnpm tsgo --noEmit`
+- Run `pnpm lint:js --quiet`
+- Run `pnpm build`
+- Find if the file has tests for it, or if there's `__test__` folder or the parent folder has tests, and run. 
+- Fixed ALL resulting errors
+
+## Context Management
+
+5. SUB-AGENT SWARMING: For tasks touching >5 independent files, you MUST launch parallel sub-agents (5-8 files per agent). Each agent gets its own context window. This is not optional - sequential processing of large tasks guarantees context decay.
+
+6. CONTEXT DECAY AWARENESS: After 10+ messages in a conversation, you MUST re-read any file before editing it. Do not trust your memory of file contents. Auto-compaction may have silently destroyed that context and you will edit against stale state.
+
+7. FILE READ BUDGET: Each file read is capped at 2,000 lines. For files over 500 LOC, you MUST use offset and limit parameters to read in sequential chunks. Never assume you have seen a complete file from a single read.
+
+8. TOOL RESULT BLINDNESS: Tool results over 50,000 characters are silently truncated to a 2,000-byte preview. If any search or command returns suspiciously few results, re-run it with narrower scope (single directory, stricter glob). State when you suspect truncation occu>
+
+## Edit Safety
+
+9.  EDIT INTEGRITY: Before EVERY file edit, re-read the file. After editing, read it again to confirm the change applied correctly. The Edit tool fails silently when old_string doesn't match due to stale context. Never batch more than 3 edits to the same file without a ve>
+
+10. NO SEMANTIC SEARCH: You have grep, not an AST. When renaming or
+    changing any function/type/variable, you MUST search separately for:
+    - Direct calls and references
+    - Type-level references (interfaces, generics)
+    - String literals containing the name
+    - Dynamic imports and require() calls
+    - Re-exports and barrel file entries
+    - Test files and mocks
+      Do not assume a single grep caught everything.

frontend/CLAUDE.md

@@ -0,0 +1 @@
+AGENTS.md

frontend/public/locales/en-GB/titles.json

@@ -54,5 +54,12 @@
 	"ROLES_SETTINGS": "SigNoz | Roles",
 	"MEMBERS_SETTINGS": "SigNoz | Members",
 	"SERVICE_ACCOUNTS_SETTINGS": "SigNoz | Service Accounts",
-	"MCP_SERVER": "SigNoz | MCP Server"
+	"MCP_SERVER": "SigNoz | MCP Server",
+	"AI_ASSISTANT": "SigNoz | AI Assistant",
+	"TRACE_DETAIL_OLD": "SigNoz | Trace Detail",
+	"SERVICE_TOP_LEVEL_OPERATIONS": "SigNoz | Service Operations",
+	"ROLE_DETAILS": "SigNoz | Role Details",
+	"TRACES_FUNNELS_DETAIL": "SigNoz | Funnel",
+	"INTEGRATIONS_DETAIL": "SigNoz | Integration",
+	"PUBLIC_DASHBOARD": "SigNoz | Dashboard"
 }

frontend/public/locales/en/titles.json

@@ -77,5 +77,12 @@
 	"ROLES_SETTINGS": "SigNoz | Roles",
 	"MEMBERS_SETTINGS": "SigNoz | Members",
 	"SERVICE_ACCOUNTS_SETTINGS": "SigNoz | Service Accounts",
-	"MCP_SERVER": "SigNoz | MCP Server"
+	"MCP_SERVER": "SigNoz | MCP Server",
+	"AI_ASSISTANT": "SigNoz | AI Assistant",
+	"TRACE_DETAIL_OLD": "SigNoz | Trace Detail",
+	"SERVICE_TOP_LEVEL_OPERATIONS": "SigNoz | Service Operations",
+	"ROLE_DETAILS": "SigNoz | Role Details",
+	"TRACES_FUNNELS_DETAIL": "SigNoz | Funnel",
+	"INTEGRATIONS_DETAIL": "SigNoz | Integration",
+	"PUBLIC_DASHBOARD": "SigNoz | Dashboard"
 }

frontend/src/api/generated/services/sigNoz.schemas.ts

@@ -2051,6 +2051,19 @@ export interface ErrorsResponseerroradditionalDTO {
 	message?: string;
 }
 
+export enum ErrorsResponseretrypolicyDTO {
+	never = 'never',
+	immediate = 'immediate',
+	backoff = 'backoff',
+	after = 'after',
+	after_fix = 'after_fix',
+	after_auth = 'after_auth',
+}
+export interface ErrorsResponseretryjsonDTO {
+	delay?: TimeDurationDTO;
+	policy?: ErrorsResponseretrypolicyDTO;
+}
+
 export interface ErrorsJSONDTO {
 	/**
 	 * @type string
@@ -2060,10 +2073,23 @@ export interface ErrorsJSONDTO {
 	 * @type array
 	 */
 	errors?: ErrorsResponseerroradditionalDTO[];
+	/**
+	 * @type array
+	 */
+	invalidReferences?: string[];
 	/**
 	 * @type string
 	 */
 	message: string;
+	retry?: ErrorsResponseretryjsonDTO;
+	/**
+	 * @type array
+	 */
+	suggestions?: string[];
+	/**
+	 * @type string
+	 */
+	type?: string;
 	/**
 	 * @type string
 	 */

frontend/src/components/cmdKPalette/cmdKPalette.tsx

@@ -1,4 +1,5 @@
 import React, { useEffect } from 'react';
+import { useLocation } from 'react-router-dom';
 import {
 	CommandDialog,
 	CommandEmpty,
@@ -9,7 +10,17 @@ import {
 	CommandShortcut,
 } from '@signozhq/ui/command';
 import logEvent from 'api/common/logEvent';
+import {
+	AIAssistantEvents,
+	AIAssistantOpenSource,
+} from 'container/AIAssistant/events';
+import { normalizePage } from 'container/AIAssistant/hooks/useAIAssistantAnalyticsContext';
+import {
+	openAIAssistantModal,
+	useAIAssistantStore,
+} from 'container/AIAssistant/store/useAIAssistantStore';
 import { useThemeMode } from 'hooks/useDarkMode';
+import { useIsAIAssistantEnabled } from 'hooks/useIsAIAssistantEnabled';
 import history from 'lib/history';
 import { ROLES as UserRole } from 'types/roles';
 
@@ -37,6 +48,11 @@ export function CmdKPalette({
 	const { open, setOpen } = useCmdK();
 
 	const { setAutoSwitch, setTheme, theme } = useThemeMode();
+	const location = useLocation();
+	const isAIAssistantEnabled = useIsAIAssistantEnabled();
+	const startNewConversation = useAIAssistantStore(
+		(s) => s.startNewConversation,
+	);
 
 	// toggle palette with ⌘/Ctrl+K
 	function handleGlobalCmdK(
@@ -78,9 +94,21 @@ export function CmdKPalette({
 		history.push(key);
 	}
 
+	const handleOpenAIAssistant = (): void => {
+		void logEvent(AIAssistantEvents.Opened, {
+			source: AIAssistantOpenSource.Cmdk,
+			currentPage: normalizePage(location.pathname),
+		});
+		startNewConversation();
+		openAIAssistantModal();
+	};
+
 	const actions = createShortcutActions({
 		navigate: onClickHandler,
 		handleThemeChange,
+		aiAssistant: isAIAssistantEnabled
+			? { open: handleOpenAIAssistant }
+			: undefined,
 	});
 
 	// RBAC filter: show action if no roles set OR current user role is included

frontend/src/constants/shortcutActions.tsx

@@ -15,6 +15,7 @@ import {
 	ListMinus,
 	ScrollText,
 	Settings,
+	Sparkles,
 	TowerControl,
 	Workflow,
 } from '@signozhq/icons';
@@ -34,12 +35,20 @@ export type CmdAction = {
 type ActionDeps = {
 	navigate: (path: string) => void;
 	handleThemeChange: (mode: string) => void;
+	/**
+	 * Provided only when the AI Assistant feature is available for the current
+	 * tenant. When present, the palette surfaces an "Open AI Assistant" entry
+	 * at the top; when absent, the action is omitted entirely.
+	 */
+	aiAssistant?: {
+		open: () => void;
+	};
 };
 
 export function createShortcutActions(deps: ActionDeps): CmdAction[] {
-	const { navigate, handleThemeChange } = deps;
+	const { navigate, handleThemeChange, aiAssistant } = deps;
 
-	return [
+	const actions: CmdAction[] = [
 		{
 			id: 'home',
 			name: 'Go to Home',
@@ -279,4 +288,19 @@ export function createShortcutActions(deps: ActionDeps): CmdAction[] {
 			perform: (): void => navigate(ROUTES.MEMBERS_SETTINGS),
 		},
 	];
+
+	if (aiAssistant) {
+		actions.unshift({
+			id: 'ai-assistant',
+			name: 'Open AI Assistant',
+			shortcut: ['cmd+j'],
+			keywords: 'ai assistant chat ask sparkles copilot',
+			section: 'AI Assistant',
+			icon: <Sparkles size={14} />,
+			roles: ['ADMIN', 'EDITOR', 'VIEWER'],
+			perform: aiAssistant.open,
+		});
+	}
+
+	return actions;
 }

frontend/src/container/AIAssistant/AIAssistantModal/AIAssistantModal.tsx

@@ -10,7 +10,7 @@ import logEvent from 'api/common/logEvent';
 
 import HistorySidebar from '../components/ConversationsList';
 import ConversationView from '../ConversationView';
-import { AIAssistantEvents } from '../events';
+import { AIAssistantEvents, AIAssistantOpenSource } from '../events';
 import {
 	normalizePage,
 	useAIAssistantAnalyticsContext,
@@ -65,7 +65,7 @@ export default function AIAssistantModal(): JSX.Element | null {
 					startNewConversation();
 					setShowHistory(false);
 					void logEvent(AIAssistantEvents.Opened, {
-						source: 'shortcut',
+						source: AIAssistantOpenSource.Shortcut,
 						currentPage: normalizePage(pathname),
 					});
 					openModal();
@@ -162,57 +162,57 @@ export default function AIAssistantModal(): JSX.Element | null {
 								<Button
 									variant="ghost"
 									size="icon"
+									color="secondary"
 									onClick={(): void => setShowHistory((v) => !v)}
 									aria-label="Toggle conversations"
 									className={showHistory ? styles.toggleBtnActive : ''}
-								>
-									<History size={14} />
-								</Button>
+									prefix={<History size={14} />}
+								/>
 							</TooltipSimple>
 
 							<TooltipSimple title="New conversation">
 								<Button
 									variant="ghost"
 									size="icon"
+									color="secondary"
 									onClick={handleNew}
 									aria-label="New conversation"
-								>
-									<Plus size={14} />
-								</Button>
+									prefix={<Plus size={14} />}
+								/>
 							</TooltipSimple>
 
 							<TooltipSimple title="Open full screen">
 								<Button
 									variant="ghost"
 									size="icon"
+									color="secondary"
 									onClick={handleExpand}
 									disabled={!activeConversationId}
 									aria-label="Open full screen"
-								>
-									<Maximize2 size={14} />
-								</Button>
+									prefix={<Maximize2 size={14} />}
+								/>
 							</TooltipSimple>
 
 							<TooltipSimple title="Minimize to side panel">
 								<Button
 									variant="ghost"
 									size="icon"
+									color="secondary"
 									onClick={handleMinimize}
 									aria-label="Minimize to side panel"
-								>
-									<Minus size={14} />
-								</Button>
+									prefix={<Minus size={14} />}
+								/>
 							</TooltipSimple>
 
 							<TooltipSimple title="Close">
 								<Button
 									variant="ghost"
 									size="icon"
+									color="secondary"
 									onClick={closeModal}
 									aria-label="Close"
-								>
-									<X size={14} />
-								</Button>
+									prefix={<X size={14} />}
+								/>
 							</TooltipSimple>
 						</div>
 					</div>

frontend/src/container/AIAssistant/AIAssistantPanel/AIAssistantPanel.tsx

@@ -150,9 +150,8 @@ export default function AIAssistantPanel(): JSX.Element | null {
 								color="secondary"
 								onClick={(): void => setShowHistory((v) => !v)}
 								aria-label="Toggle conversations"
-							>
-								<History size={14} />
-							</Button>
+								prefix={<History size={14} />}
+							/>
 						</TooltipSimple>
 
 						<TooltipSimple title="New conversation">
@@ -162,9 +161,8 @@ export default function AIAssistantPanel(): JSX.Element | null {
 								color="secondary"
 								onClick={handleNew}
 								aria-label="New conversation"
-							>
-								<Plus size={14} />
-							</Button>
+								prefix={<Plus size={14} />}
+							/>
 						</TooltipSimple>
 
 						<TooltipSimple title="Open full screen">
@@ -175,9 +173,8 @@ export default function AIAssistantPanel(): JSX.Element | null {
 								onClick={handleExpand}
 								disabled={!activeConversationId}
 								aria-label="Open full screen"
-							>
-								<Maximize2 size={14} />
-							</Button>
+								prefix={<Maximize2 size={14} />}
+							/>
 						</TooltipSimple>
 
 						<TooltipSimple title="Close">
@@ -187,9 +184,8 @@ export default function AIAssistantPanel(): JSX.Element | null {
 								color="secondary"
 								onClick={closeDrawer}
 								aria-label="Close panel"
-							>
-								<X size={14} />
-							</Button>
+								prefix={<X size={14} />}
+							/>
 						</TooltipSimple>
 					</div>
 				</div>

frontend/src/container/AIAssistant/AIAssistantTrigger/AIAssistantTrigger.tsx

@@ -6,7 +6,7 @@ import logEvent from 'api/common/logEvent';
 import ROUTES from 'constants/routes';
 import { Bot } from '@signozhq/icons';
 
-import { AIAssistantEvents } from '../events';
+import { AIAssistantEvents, AIAssistantOpenSource } from '../events';
 import { normalizePage } from '../hooks/useAIAssistantAnalyticsContext';
 import {
 	openAIAssistant,
@@ -31,7 +31,7 @@ export default function AIAssistantTrigger(): JSX.Element | null {
 
 	const handleOpen = useCallback((): void => {
 		void logEvent(AIAssistantEvents.Opened, {
-			source: 'icon',
+			source: AIAssistantOpenSource.Icon,
 			currentPage: normalizePage(pathname),
 		});
 		openAIAssistant();

frontend/src/container/AIAssistant/ConversationView/ConversationView.tsx

@@ -159,6 +159,7 @@ export default function ConversationView({
 				<ConversationSkeleton />
 				<div className={inputWrapperClass}>
 					<ChatInput
+						key={conversationId}
 						onSend={handleSend}
 						disabled
 						autoContexts={autoContexts}
@@ -172,6 +173,7 @@ export default function ConversationView({
 	return (
 		<div className={styles.conversation}>
 			<VirtualizedMessages
+				key={conversationId}
 				conversationId={conversationId}
 				messages={messages}
 				isStreaming={isStreamingHere}
@@ -184,6 +186,7 @@ export default function ConversationView({
 			)}
 			<div className={inputWrapperClass}>
 				<ChatInput
+					key={conversationId}
 					onSend={handleSend}
 					onCancel={handleCancel}
 					disabled={inputDisabled}

frontend/src/container/AIAssistant/components/ApprovalCard/ApprovalCard.tsx

@@ -11,6 +11,7 @@ import {
 	DialogTitle,
 } from '@signozhq/ui/dialog';
 import { ToggleGroup, ToggleGroupItem } from '@signozhq/ui/toggle-group';
+import { TooltipSimple } from '@signozhq/ui/tooltip';
 import type {
 	ApprovalEventDTO,
 	ApprovalEventDTODiff,
@@ -100,16 +101,16 @@ export default function ApprovalCard({
 				<div className={styles.diffSection}>
 					<div className={styles.diffHeader}>
 						<span className={styles.diffHeaderLabel}>Diff</span>
-						<Button
-							variant="link"
-							size="sm"
-							color="secondary"
-							onClick={(): void => setDiffExpanded(true)}
-							title="Expand diff"
-							aria-label="Expand diff"
-						>
-							<Maximize2 size={12} />
-						</Button>
+						<TooltipSimple title="Expand diff">
+							<Button
+								variant="link"
+								size="sm"
+								color="secondary"
+								onClick={(): void => setDiffExpanded(true)}
+								aria-label="Expand diff"
+								prefix={<Maximize2 size={12} />}
+							/>
+						</TooltipSimple>
 					</div>
 					<DiffView diff={approval.diff} />
 				</div>
@@ -119,6 +120,8 @@ export default function ApprovalCard({
 				<DialogContent
 					className={styles.diffDialog}
 					style={{ width: '80vw', maxWidth: '80vw', height: '70vh' }}
+					// Skip auto-focus — otherwise the first Copy button opens its tooltip on dialog open.
+					onOpenAutoFocus={(e): void => e.preventDefault()}
 				>
 					<DialogHeader>
 						<DialogTitle>Approval diff</DialogTitle>
@@ -134,19 +137,22 @@ export default function ApprovalCard({
 								size="sm"
 								value={viewMode}
 								onChange={(next): void => {
-									// Radix `single` group can emit '' when the active item
-									// is clicked again — preserve the current mode.
+									// Radix `single` group can emit '' when the active item is clicked again.
 									if (next === 'split' || next === 'unified') {
 										setViewMode(next);
 									}
 								}}
 							>
-								<ToggleGroupItem value="split" aria-label="Split view">
-									<Columns2 size={12} />
-								</ToggleGroupItem>
-								<ToggleGroupItem value="unified" aria-label="Unified view">
-									<List size={12} />
-								</ToggleGroupItem>
+								<TooltipSimple title="Split view">
+									<ToggleGroupItem value="split" aria-label="Split view">
+										<Columns2 size={12} />
+									</ToggleGroupItem>
+								</TooltipSimple>
+								<TooltipSimple title="Unified view">
+									<ToggleGroupItem value="unified" aria-label="Unified view">
+										<List size={12} />
+									</ToggleGroupItem>
+								</TooltipSimple>
 							</ToggleGroup>
 							<ToggleGroup
 								type="multiple"
@@ -154,12 +160,16 @@ export default function ApprovalCard({
 								value={wrapText ? ['wrap'] : []}
 								onChange={(next): void => setWrapText(next.includes('wrap'))}
 							>
-								<ToggleGroupItem
-									value="wrap"
-									aria-label={wrapText ? 'Disable text wrap' : 'Wrap long lines'}
+								<TooltipSimple
+									title={wrapText ? 'Disable text wrap' : 'Wrap long lines'}
 								>
-									<WrapText size={12} />
-								</ToggleGroupItem>
+									<ToggleGroupItem
+										value="wrap"
+										aria-label={wrapText ? 'Disable text wrap' : 'Wrap long lines'}
+									>
+										<WrapText size={12} />
+									</ToggleGroupItem>
+								</TooltipSimple>
 							</ToggleGroup>
 						</div>
 						{approval.diff && (
@@ -457,15 +467,16 @@ function CopyButton({ text, label }: CopyButtonProps): JSX.Element {
 	};
 
 	return (
-		<Button
-			variant="ghost"
-			size="sm"
-			color="secondary"
-			onClick={handleCopy}
-			title={copied ? `Copied ${label}` : `Copy ${label}`}
-			aria-label={copied ? `Copied ${label}` : `Copy ${label}`}
-		>
-			{copied ? <Check size={12} /> : <Copy size={12} />}
-		</Button>
+		<TooltipSimple title={copied ? `Copied ${label}` : `Copy ${label}`}>
+			<Button
+				variant="ghost"
+				size="sm"
+				color="secondary"
+				onClick={handleCopy}
+				aria-label={copied ? `Copied ${label}` : `Copy ${label}`}
+			>
+				{copied ? <Check size={12} /> : <Copy size={12} />}
+			</Button>
+		</TooltipSimple>
 	);
 }

frontend/src/container/AIAssistant/components/ChatInput/ChatInput.module.scss

@@ -8,12 +8,7 @@
 	border-radius: var(--radius-2);
 	padding: 8px;
 	border: 1px solid var(--l1-border);
-	transition: border-color 0.15s;
 	position: relative;
-
-	&:focus-within {
-		border-color: var(--l1-border);
-	}
 }
 
 .attachments {
@@ -129,6 +124,18 @@
 	border: 1px solid var(--l2-border);
 	border-radius: var(--radius-2);
 	padding: 4px;
+	transition:
+		border-color 0.15s,
+		box-shadow 0.15s;
+
+	// Scope the focus ring to the textarea row only — the surrounding
+	// chrome (context chips, "Add Context", mic, send) sits outside this
+	// element and stays unaffected when the cursor enters the textarea.
+	&:focus-within {
+		border-color: var(--accent-primary);
+		box-shadow: 0 0 0 1px
+			color-mix(in srgb, var(--accent-primary), transparent 70%);
+	}
 }
 
 .footer {
@@ -244,16 +251,24 @@
 }
 
 .contextPopoverCategoryItem {
+	// Override DS Button's centered layout.
+	--button-justify-content: flex-start;
 	display: flex;
 	align-items: center;
+	justify-content: flex-start;
 	gap: 8px;
+	width: 100%;
 	height: 32px;
 	padding: 0 8px;
-	border: 1px solid color-mix(in srgb, var(--l1-foreground), transparent 96%);
-	border-radius: var(--radius-2);
+	background: transparent;
+	color: inherit;
+	font: inherit;
 	font-size: 12px;
 	font-weight: 550;
 	text-align: left;
+	border: 1px solid color-mix(in srgb, var(--l1-foreground), transparent 96%);
+	border-radius: var(--radius-2);
+	appearance: none;
 	cursor: pointer;
 	transition:
 		background 0.15s ease,
@@ -309,17 +324,24 @@
 }
 
 .contextPopoverEntityItem {
+	// Override DS Button's centered layout.
+	--button-justify-content: flex-start;
 	display: flex;
 	align-items: center;
+	justify-content: flex-start;
 	gap: 10px;
+	width: 100%;
 	padding: 8px;
-	border: 1px solid color-mix(in srgb, var(--l1-foreground), transparent 96%);
-	border-radius: var(--radius-2);
+	background: transparent;
 	color: var(--l1-foreground);
+	font: inherit;
 	font-size: 12px;
 	font-weight: 500;
 	line-height: 1.35;
 	text-align: left;
+	border: 1px solid color-mix(in srgb, var(--l1-foreground), transparent 96%);
+	border-radius: var(--radius-2);
+	appearance: none;
 	cursor: pointer;
 	// Required for the inner span's `text-overflow: ellipsis` to engage —
 	// flex items default to `min-width: auto` (intrinsic width) and would
@@ -385,6 +407,11 @@
 	border-radius: 50%;
 	border: none;
 	cursor: pointer;
+	// Reset native <button> defaults so the 24px circle isn't inflated by
+	// browser-default padding / font metrics.
+	padding: 0;
+	font: inherit;
+	appearance: none;
 }
 
 .micDiscard {

frontend/src/container/AIAssistant/components/ChatInput/ChatInput.tsx

@@ -1,4 +1,10 @@
-import { useCallback, useEffect, useRef, useState } from 'react';
+import {
+	useCallback,
+	useEffect,
+	useLayoutEffect,
+	useRef,
+	useState,
+} from 'react';
 import { useQueryClient } from 'react-query';
 import cx from 'classnames';
 import { Badge } from '@signozhq/ui/badge';
@@ -26,7 +32,11 @@ import { useSelector } from 'react-redux';
 import { AppState } from 'store/reducers';
 import { GlobalReducer } from 'types/reducer/globalTime';
 
-import { AIAssistantEvents, getBrowserInfo } from '../../events';
+import {
+	AIAssistantEvents,
+	VoiceInputSource,
+	getBrowserInfo,
+} from '../../events';
 import { useAIAssistantAnalyticsContext } from '../../hooks/useAIAssistantAnalyticsContext';
 import { useSpeechRecognition } from '../../hooks/useSpeechRecognition';
 import { MessageAttachment } from '../../types';
@@ -142,6 +152,10 @@ function autoContextCategory(ctx: MessageContext): string {
 
 const MAX_INPUT_LENGTH = 20000;
 const WARNING_THRESHOLD = 15000;
+// Cap for the auto-growing composer. Past this, the textarea stops growing
+// and starts scrolling internally so the message list above doesn't get
+// squeezed in tighter container variants (e.g. the floating panel).
+const TEXTAREA_MAX_HEIGHT_PX = 200;
 const HOME_SERVICES_INTERVAL = 30 * 60 * 1000;
 /** sessionStorage key for the "voice input failed this tab" flag. */
 const VOICE_UNAVAILABLE_KEY = 'ai-assistant-voice-unavailable';
@@ -224,6 +238,18 @@ export default function ChatInput({
 	const [activeContextCategory, setActiveContextCategory] =
 		useState<ContextCategory>('Dashboards');
 	const [pickerSearchQuery, setPickerSearchQuery] = useState('');
+	// Refs to each category tab so we can move DOM focus to the newly-active
+	// tab on ArrowUp/ArrowDown. Without this the roving-tabindex pattern
+	// stalls: focus stays on the original button (whose closure has the old
+	// category), so subsequent arrow keys never advance past the second tab.
+	const categoryTabRefs = useRef(
+		new Map<ContextCategory, HTMLButtonElement | null>(),
+	);
+	// Refs to each entity row in the active tab panel, so we can cross from
+	// the category tablist (ArrowRight) into the panel and step through
+	// entities with ArrowUp/Down. Array is rewritten each render — there's
+	// only ever one tab panel mounted so stale indices clear naturally.
+	const entityRefs = useRef<(HTMLButtonElement | null)[]>([]);
 	const queryClient = useQueryClient();
 
 	// When the picker was opened by typing `@` in the textarea, this holds the
@@ -303,11 +329,92 @@ export default function ChatInput({
 		[mentionRange, selectedContexts, text],
 	);
 
+	const focusCategory = useCallback((category: ContextCategory) => {
+		setActiveContextCategory(category);
+		setPickerSearchQuery('');
+		categoryTabRefs.current.get(category)?.focus();
+	}, []);
+
+	const handleCategoryKeyDown = useCallback(
+		(
+			e: React.KeyboardEvent<HTMLButtonElement>,
+			category: ContextCategory,
+		): void => {
+			const total = CONTEXT_CATEGORIES.length;
+			const idx = CONTEXT_CATEGORIES.indexOf(category);
+			if (e.key === 'ArrowDown') {
+				e.preventDefault();
+				focusCategory(CONTEXT_CATEGORIES[(idx + 1) % total]);
+			} else if (e.key === 'ArrowUp') {
+				e.preventDefault();
+				focusCategory(CONTEXT_CATEGORIES[(idx - 1 + total) % total]);
+			} else if (e.key === 'Home') {
+				e.preventDefault();
+				focusCategory(CONTEXT_CATEGORIES[0]);
+			} else if (e.key === 'End') {
+				e.preventDefault();
+				focusCategory(CONTEXT_CATEGORIES[total - 1]);
+			} else if (e.key === 'ArrowRight') {
+				// Cross from tablist into entity panel.
+				e.preventDefault();
+				entityRefs.current[0]?.focus();
+			}
+		},
+		[focusCategory],
+	);
+
+	const handleEntityKeyDown = useCallback(
+		(e: React.KeyboardEvent<HTMLButtonElement>, index: number): void => {
+			const count = entityRefs.current.length;
+			if (count === 0) {
+				return;
+			}
+			const focusAt = (i: number): void => {
+				e.preventDefault();
+				entityRefs.current[i]?.focus();
+			};
+			switch (e.key) {
+				case 'ArrowDown':
+					focusAt((index + 1) % count);
+					break;
+				case 'ArrowUp':
+					focusAt((index - 1 + count) % count);
+					break;
+				case 'Home':
+					focusAt(0);
+					break;
+				case 'End':
+					focusAt(count - 1);
+					break;
+				case 'ArrowLeft':
+					// Cross back to tablist.
+					e.preventDefault();
+					categoryTabRefs.current.get(activeContextCategory)?.focus();
+					break;
+				default:
+			}
+		},
+		[activeContextCategory],
+	);
+
 	// Focus the textarea when this component mounts (panel/modal open)
 	useEffect(() => {
 		textareaRef.current?.focus();
 	}, []);
 
+	// Auto-grow the textarea so long prompts aren't trapped in a 2-line
+	// scrolling porthole. Reset to `auto` first to let the field shrink back
+	// down when the user deletes content, then snap to scrollHeight capped at
+	// TEXTAREA_MAX_HEIGHT_PX (overflow-y: auto in CSS handles the rest).
+	useLayoutEffect(() => {
+		const el = textareaRef.current;
+		if (!el) {
+			return;
+		}
+		el.style.height = 'auto';
+		el.style.height = `${Math.min(el.scrollHeight, TEXTAREA_MAX_HEIGHT_PX)}px`;
+	}, [text]);
+
 	const handleSend = useCallback(async () => {
 		const trimmed = text.trim();
 		if (!trimmed && pendingFiles.length === 0) {
@@ -382,7 +489,7 @@ export default function ChatInput({
 	// start time so we can attribute `durationMs` on the Voice input used
 	// event regardless of which control ended the session.
 	const voiceStartedAtRef = useRef<number | null>(null);
-	const voiceSourceRef = useRef<'button' | 'shortcut' | null>(null);
+	const voiceSourceRef = useRef<VoiceInputSource | null>(null);
 	// Set to true after a `network`, `not-allowed`, or `not-supported` failure
 	// so we hide the mic button for the rest of the tab session — silent
 	// retries don't help, and Chromium derivatives without the Google Speech
@@ -459,7 +566,7 @@ export default function ChatInput({
 	const showMic = isSupported && micPermission !== 'denied' && !voiceUnavailable;
 
 	const startVoiceInput = useCallback(
-		(source: 'button' | 'shortcut') => {
+		(source: VoiceInputSource) => {
 			// Defense in depth: the button is hidden when `voiceUnavailable` is
 			// true, but the PTT shortcut listener can still call us. Bailing here
 			// keeps a single source of truth and prevents repeat `Voice input
@@ -536,7 +643,7 @@ export default function ChatInput({
 				return; // ignore auto-repeat
 			}
 			pttActiveRef.current = true;
-			startVoiceInput('shortcut');
+			startVoiceInput(VoiceInputSource.Shortcut);
 		};
 
 		const handleKeyUp = (e: KeyboardEvent): void => {
@@ -724,6 +831,12 @@ export default function ChatInput({
 				entity.value.toLowerCase().includes(activeQuery),
 			)
 		: contextEntitiesByCategory[activeContextCategory];
+	// Truncate the ref array to match the current entity count so that
+	// switching from a large category (e.g. 100 dashboards) to a smaller one
+	// doesn't leave stale `null` slots from earlier renders. Keyboard nav math
+	// already uses `filteredContextOptions.length` for the modulo, so stale
+	// slots wouldn't be reached — this is purely housekeeping.
+	entityRefs.current.length = filteredContextOptions.length;
 	const { isLoading: isActiveContextLoading, isError: isActiveContextError } =
 		contextCategoryStateByCategory[activeContextCategory];
 	const currentLength = text.length;
@@ -830,7 +943,7 @@ export default function ChatInput({
 					onKeyDown={handleKeyDown}
 					disabled={disabled}
 					maxLength={MAX_INPUT_LENGTH}
-					rows={2}
+					rows={3}
 				/>
 			</div>
 			{showTextWarning && (
@@ -877,15 +990,37 @@ export default function ChatInput({
 							sideOffset={8}
 						>
 							<div className={styles.contextPopoverContent}>
-								<div className={styles.contextPopoverCategories}>
+								<div
+									className={styles.contextPopoverCategories}
+									role="tablist"
+									aria-orientation="vertical"
+									aria-label="Context categories"
+								>
 									{CONTEXT_CATEGORIES.map((category) => {
 										const CategoryIcon = CONTEXT_CATEGORY_ICONS[category];
 										const isActive = activeContextCategory === category;
 										return (
-											<div
+											<Button
 												key={category}
+												ref={(el): void => {
+													categoryTabRefs.current.set(category, el);
+												}}
+												type="button"
+												variant="ghost"
+												color="secondary"
+												size="sm"
 												role="tab"
-												tabIndex={0}
+												id={`ai-context-tab-${category}`}
+												// Single stable panel id shared by every tab: only the
+												// active category's tabpanel is rendered, so per-category
+												// `aria-controls` ids would point at nonexistent nodes
+												// for the two inactive tabs. APG allows a single
+												// dynamic panel whose `aria-labelledby` swaps to the
+												// active tab.
+												aria-controls="ai-context-tabpanel"
+												// Roving tabindex: only the active tab participates in
+												// the Tab sequence; arrow keys move between tabs.
+												tabIndex={isActive ? 0 : -1}
 												aria-selected={isActive}
 												className={cx(styles.contextPopoverCategoryItem, {
 													[styles.active]: isActive,
@@ -894,22 +1029,21 @@ export default function ChatInput({
 													setActiveContextCategory(category);
 													setPickerSearchQuery('');
 												}}
-												onKeyDown={(e): void => {
-													if (e.key === 'Enter' || e.key === ' ') {
-														e.preventDefault();
-														setActiveContextCategory(category);
-														setPickerSearchQuery('');
-													}
-												}}
+												onKeyDown={(e): void => handleCategoryKeyDown(e, category)}
+												prefix={<CategoryIcon size={13} />}
 											>
-												<CategoryIcon size={13} />
 												<span>{category}</span>
-											</div>
+											</Button>
 										);
 									})}
 								</div>
 
-								<div className={styles.contextPopoverRight}>
+								<div
+									className={styles.contextPopoverRight}
+									role="tabpanel"
+									id="ai-context-tabpanel"
+									aria-labelledby={`ai-context-tab-${activeContextCategory}`}
+								>
 									<div className={styles.contextPopoverSearch}>
 										<Input
 											type="text"
@@ -939,7 +1073,7 @@ export default function ChatInput({
 												No matching entities
 											</div>
 										) : (
-											filteredContextOptions.map((option) => {
+											filteredContextOptions.map((option, index) => {
 												const isSelected = selectedContexts.some(
 													(item) =>
 														item.category === activeContextCategory &&
@@ -947,8 +1081,16 @@ export default function ChatInput({
 												);
 
 												return (
-													<div
+													<Button
 														key={option.id}
+														ref={(el): void => {
+															entityRefs.current[index] = el;
+														}}
+														type="button"
+														variant="ghost"
+														color="secondary"
+														size="sm"
+														aria-pressed={isSelected}
 														className={cx(styles.contextPopoverEntityItem, {
 															[styles.selected]: isSelected,
 														})}
@@ -959,11 +1101,12 @@ export default function ChatInput({
 																option.value,
 															)
 														}
+														onKeyDown={(e): void => handleEntityKeyDown(e, index)}
 													>
 														<span className={styles.contextPopoverEntityItemText}>
 															{option.value}
 														</span>
-													</div>
+													</Button>
 												);
 											})
 										)}
@@ -977,14 +1120,24 @@ export default function ChatInput({
 				<div className={styles.rightActions}>
 					{showMic &&
 						(isListening ? (
-							<div className={styles.micRecording}>
-								<div
-									className={cx(styles.micDiscard, styles.secondary)}
-									onClick={handleDiscard}
-									aria-label="Discard recording"
-								>
-									<X size={12} />
-								</div>
+							<div
+								className={styles.micRecording}
+								role="status"
+								aria-live="polite"
+								aria-label="Recording voice input"
+							>
+								<TooltipSimple title="Discard recording">
+									<Button
+										type="button"
+										variant="ghost"
+										size="icon"
+										color="secondary"
+										className={cx(styles.micDiscard, styles.secondary)}
+										onClick={handleDiscard}
+										aria-label="Discard recording"
+										prefix={<X size={12} />}
+									/>
+								</TooltipSimple>
 								<span className={styles.micWaves} aria-hidden="true">
 									<span />
 									<span />
@@ -995,26 +1148,30 @@ export default function ChatInput({
 									<span />
 									<span />
 								</span>
-								<div
-									className={cx(styles.micStop, styles.destructive)}
-									onClick={handleStopAndSend}
-									aria-label="Stop and send"
-								>
-									<Square size={9} fill="currentColor" strokeWidth={0} />
-								</div>
+								<TooltipSimple title="Stop and send">
+									<Button
+										type="button"
+										variant="ghost"
+										size="icon"
+										color="destructive"
+										className={cx(styles.micStop, styles.destructive)}
+										onClick={handleStopAndSend}
+										aria-label="Stop and send"
+										prefix={<Square size={9} fill="currentColor" strokeWidth={0} />}
+									/>
+								</TooltipSimple>
 							</div>
 						) : (
 							<TooltipSimple title="Voice input">
 								<Button
 									variant="ghost"
 									size="icon"
-									onClick={(): void => startVoiceInput('button')}
+									onClick={(): void => startVoiceInput(VoiceInputSource.Button)}
 									disabled={disabled}
 									aria-label="Start voice input"
 									className={styles.micBtn}
-								>
-									<Mic size={14} />
-								</Button>
+									prefix={<Mic size={14} />}
+								/>
 							</TooltipSimple>
 						))}
 
@@ -1026,21 +1183,21 @@ export default function ChatInput({
 								color="destructive"
 								onClick={onCancel}
 								aria-label="Stop generating"
-							>
-								<Square size={10} fill="currentColor" strokeWidth={0} />
-							</Button>
+								prefix={<Square size={10} fill="currentColor" strokeWidth={0} />}
+							/>
 						</TooltipSimple>
 					) : (
-						<Button
-							variant="solid"
-							size="icon"
-							color="primary"
-							onClick={isListening ? handleStopAndSend : handleSend}
-							disabled={disabled || (!text.trim() && pendingFiles.length === 0)}
-							aria-label="Send message"
-						>
-							<Send size={14} />
-						</Button>
+						<TooltipSimple title="Send message">
+							<Button
+								variant="solid"
+								size="icon"
+								color="primary"
+								onClick={isListening ? handleStopAndSend : handleSend}
+								disabled={disabled || (!text.trim() && pendingFiles.length === 0)}
+								aria-label="Send message"
+								prefix={<Send size={14} />}
+							/>
+						</TooltipSimple>
 					)}
 				</div>
 			</div>

frontend/src/container/AIAssistant/components/ClarificationForm/ClarificationForm.module.scss

@@ -64,6 +64,19 @@
 	gap: 4px;
 }
 
+// Mirrors `.field` for the multi_select group, but resets the browser's
+// default `<fieldset>` border/padding/margin so the visual matches the
+// `<div>`-based field rows.
+.multiSelectFieldset {
+	display: flex;
+	flex-direction: column;
+	gap: 4px;
+	margin: 0;
+	padding: 0;
+	border: 0;
+	min-width: 0;
+}
+
 .label {
 	font-size: 12px;
 	font-weight: 500;

frontend/src/container/AIAssistant/components/ClarificationForm/ClarificationForm.tsx

@@ -63,7 +63,14 @@ export default function ClarificationForm({
 		setAnswers((prev) => ({ ...prev, [id]: value }));
 	};
 
+	const isFormValid = fields.every(
+		(f) => !f.required || isFieldFilled(f, answers[f.id]),
+	);
+
 	const handleSubmit = async (): Promise<void> => {
+		if (!isFormValid) {
+			return;
+		}
 		setSubmitted(true);
 		// Approximate queryLength as the JSON encoding of the form answers — the
 		// clarification API doesn't render a single user-visible string, but the
@@ -136,7 +143,7 @@ export default function ClarificationForm({
 					variant="solid"
 					color="primary"
 					onClick={handleSubmit}
-					disabled={isStreaming}
+					disabled={isStreaming || !isFormValid}
 					prefix={<Send />}
 				>
 					Submit
@@ -162,8 +169,9 @@ export default function ClarificationForm({
 
 /**
  * Per-type seed value. The DTO's `default` is `string | string[] | null`,
- * which doesn't fit boolean fields cleanly — we coerce 'true'/'false' strings
- * for them, fall back to `[]` for multi_select, and the raw string otherwise.
+ * which doesn't fit boolean / number fields cleanly — we coerce 'true'/'false'
+ * strings for booleans, parse number defaults out of the string form,
+ * fall back to `[]` for multi_select, and the raw string otherwise.
  */
 function initialAnswerFor(f: ClarificationFieldEventDTO): unknown {
 	const raw = f.default;
@@ -175,9 +183,41 @@ function initialAnswerFor(f: ClarificationFieldEventDTO): unknown {
 	if (f.type === ClarificationFieldTypeDTO.multi_select) {
 		return Array.isArray(raw) ? raw : [];
 	}
+	if (f.type === ClarificationFieldTypeDTO.number) {
+		// Server sends number defaults as strings (e.g. `"5"`). Parse so the
+		// stored value is a real `number` — otherwise `isFieldFilled` (which
+		// requires `typeof === 'number'`) rejects a visibly-filled field and
+		// Submit stays disabled.
+		if (typeof raw !== 'string' || raw === '') {
+			return null;
+		}
+		const parsed = Number(raw);
+		return Number.isNaN(parsed) ? null : parsed;
+	}
 	return raw ?? '';
 }
 
+// Whether a required field has been answered. Booleans are always considered
+// filled (they're initialised to a concrete true/false). For other types we
+// reject empty strings, empty arrays, NaN numbers, and `null` (which the
+// number input emits when its raw value is `''` — `Number('')` would
+// otherwise silently coerce to `0` and read as a valid answer).
+function isFieldFilled(
+	field: ClarificationFieldEventDTO,
+	value: unknown,
+): boolean {
+	switch (field.type) {
+		case ClarificationFieldTypeDTO.multi_select:
+			return Array.isArray(value) && value.length > 0;
+		case ClarificationFieldTypeDTO.boolean:
+			return true;
+		case ClarificationFieldTypeDTO.number:
+			return typeof value === 'number' && !Number.isNaN(value);
+		default:
+			return typeof value === 'string' && value.trim().length > 0;
+	}
+}
+
 interface FieldInputProps {
 	field: ClarificationFieldEventDTO;
 	value: unknown;
@@ -216,13 +256,21 @@ function FieldInput({ field, value, onChange }: FieldInputProps): JSX.Element {
 			<div className={styles.field}>
 				<label className={styles.label} htmlFor={id}>
 					{label}
-					{required && <span className={styles.required}>*</span>}
+					{required && (
+						<span className={styles.required} aria-hidden="true">
+							*
+						</span>
+					)}
 				</label>
 				<Select
 					value={isCustom ? CUSTOM_OPTION_SENTINEL : String(value ?? '')}
 					onChange={handleSelectChange}
 				>
-					<SelectTrigger id={id} placeholder="Select…" />
+					<SelectTrigger
+						id={id}
+						placeholder="Select…"
+						aria-required={required || undefined}
+					/>
 					{/* Pin the dropdown width to the trigger via Radix's
 					    `--radix-select-trigger-width`; otherwise the popover
 					    sizes to its widest item and looks misaligned. */}
@@ -267,7 +315,11 @@ function FieldInput({ field, value, onChange }: FieldInputProps): JSX.Element {
 					onChange={(): void => onChange(!checked)}
 				>
 					{label}
-					{required && <span className={styles.required}>*</span>}
+					{required && (
+						<span className={styles.required} aria-hidden="true">
+							*
+						</span>
+					)}
 				</Checkbox>
 			</div>
 		);
@@ -312,11 +364,21 @@ function FieldInput({ field, value, onChange }: FieldInputProps): JSX.Element {
 		};
 
 		return (
-			<div className={styles.field}>
-				<span className={styles.label}>
+			// `fieldset` + `legend` is the WCAG-recommended grouping for
+			// related checkboxes (1.3.1). SRs announce the legend before each
+			// option, so users hear the group label as context.
+			<fieldset
+				className={styles.multiSelectFieldset}
+				aria-required={required || undefined}
+			>
+				<legend className={styles.label}>
 					{label}
-					{required && <span className={styles.required}>*</span>}
-				</span>
+					{required && (
+						<span className={styles.required} aria-hidden="true">
+							*
+						</span>
+					)}
+				</legend>
 				<div className={styles.checkboxGroup}>
 					{options?.map((opt) => (
 						<Checkbox
@@ -347,7 +409,7 @@ function FieldInput({ field, value, onChange }: FieldInputProps): JSX.Element {
 						onChange={(e): void => updateCustomValue(e.target.value)}
 					/>
 				)}
-			</div>
+			</fieldset>
 		);
 	}
 
@@ -356,16 +418,29 @@ function FieldInput({ field, value, onChange }: FieldInputProps): JSX.Element {
 		<div className={styles.field}>
 			<label className={styles.label} htmlFor={id}>
 				{label}
-				{required && <span className={styles.required}>*</span>}
+				{required && (
+					<span className={styles.required} aria-hidden="true">
+						*
+					</span>
+				)}
 			</label>
 			<Input
 				id={id}
 				type={type === 'number' ? 'number' : 'text'}
 				className={styles.input}
 				value={String(value ?? '')}
-				onChange={(e): void =>
-					onChange(type === 'number' ? Number(e.target.value) : e.target.value)
-				}
+				aria-required={required || undefined}
+				onChange={(e): void => {
+					if (type === 'number') {
+						const raw = e.target.value;
+						// Map empty input to `null` instead of `Number('') === 0`
+						// so a required numeric field cleared after typing doesn't
+						// silently read as a valid `0` in `isFieldFilled`.
+						onChange(raw === '' ? null : Number(raw));
+					} else {
+						onChange(e.target.value);
+					}
+				}}
 				placeholder={label}
 			/>
 		</div>

frontend/src/container/AIAssistant/components/MessageFeedback/MessageFeedback.tsx

@@ -10,6 +10,7 @@ import { useTimezone } from 'providers/Timezone';
 
 import logEvent from 'api/common/logEvent';
 
+import { FeedbackRatingDTO } from 'api/ai-assistant/sigNozAIAssistantAPI.schemas';
 import { AIAssistantEvents } from '../../events';
 import { useAIAssistantAnalyticsContext } from '../../hooks/useAIAssistantAnalyticsContext';
 import { useAIAssistantStore } from '../../store/useAIAssistantStore';
@@ -17,6 +18,22 @@ import { FeedbackRating, Message } from '../../types';
 
 import styles from './MessageFeedback.module.scss';
 
+const FEEDBACK_ANALYTICS_RATING = {
+	[FeedbackRatingDTO.positive]: 'up',
+	[FeedbackRatingDTO.negative]: 'down',
+} as const;
+
+const VOTE_LABEL = {
+	[FeedbackRatingDTO.positive]: {
+		tooltip: 'Good response',
+		ariaLabel: 'Good response',
+	},
+	[FeedbackRatingDTO.negative]: {
+		tooltip: 'Bad response',
+		ariaLabel: 'Bad response',
+	},
+} as const;
+
 interface MessageFeedbackProps {
 	message: Message;
 	onRegenerate?: () => void;
@@ -117,7 +134,7 @@ export default function MessageFeedback({
 			if (vote === rating) {
 				return;
 			}
-			if (rating === 'negative') {
+			if (rating === FeedbackRatingDTO.negative) {
 				setNegativeComment('');
 				setIsNegativeDialogOpen(true);
 				return;
@@ -126,7 +143,7 @@ export default function MessageFeedback({
 			void logEvent(AIAssistantEvents.FeedbackSubmitted, {
 				messageId: message.id,
 				threadId,
-				rating: 'up',
+				rating: FEEDBACK_ANALYTICS_RATING[rating],
 				hasComment: false,
 				commentLength: 0,
 			});
@@ -136,17 +153,21 @@ export default function MessageFeedback({
 	);
 
 	const handleSubmitNegative = useCallback((): void => {
-		setVote('negative');
+		setVote(FeedbackRatingDTO.negative);
 		setIsNegativeDialogOpen(false);
 		const trimmed = negativeComment.trim();
 		void logEvent(AIAssistantEvents.FeedbackSubmitted, {
 			messageId: message.id,
 			threadId,
-			rating: 'down',
+			rating: FEEDBACK_ANALYTICS_RATING[FeedbackRatingDTO.negative],
 			hasComment: trimmed.length > 0,
 			commentLength: trimmed.length,
 		});
-		submitMessageFeedback(message.id, 'negative', trimmed || undefined);
+		submitMessageFeedback(
+			message.id,
+			FeedbackRatingDTO.negative,
+			trimmed || undefined,
+		);
 	}, [message.id, negativeComment, submitMessageFeedback, threadId]);
 
 	return (
@@ -160,32 +181,39 @@ export default function MessageFeedback({
 							variant="ghost"
 							onClick={handleCopy}
 							color="secondary"
+							aria-label={copied ? 'Copied' : 'Copy message'}
 						>
 							{copied ? <Check size={12} /> : <Copy size={12} />}
 						</Button>
 					</TooltipSimple>
 
-					<TooltipSimple title="Good response">
+					<TooltipSimple title={VOTE_LABEL[FeedbackRatingDTO.positive].tooltip}>
 						<Button
-							className={cx(styles.btn, { [styles.votedUp]: vote === 'positive' })}
+							className={cx(styles.btn, {
+								[styles.votedUp]: vote === FeedbackRatingDTO.positive,
+							})}
 							size="icon"
 							variant="ghost"
 							color="secondary"
-							onClick={(): void => handleVote('positive')}
+							onClick={(): void => handleVote(FeedbackRatingDTO.positive)}
+							aria-label={VOTE_LABEL[FeedbackRatingDTO.positive].ariaLabel}
+							aria-pressed={vote === FeedbackRatingDTO.positive}
 						>
 							<ThumbsUp size={12} />
 						</Button>
 					</TooltipSimple>
 
-					<TooltipSimple title="Bad response">
+					<TooltipSimple title={VOTE_LABEL[FeedbackRatingDTO.negative].tooltip}>
 						<Button
 							className={cx(styles.btn, {
-								[styles.votedDown]: vote === 'negative',
+								[styles.votedDown]: vote === FeedbackRatingDTO.negative,
 							})}
 							size="icon"
 							variant="ghost"
 							color="secondary"
-							onClick={(): void => handleVote('negative')}
+							onClick={(): void => handleVote(FeedbackRatingDTO.negative)}
+							aria-label={VOTE_LABEL[FeedbackRatingDTO.negative].ariaLabel}
+							aria-pressed={vote === FeedbackRatingDTO.negative}
 						>
 							<ThumbsDown size={12} />
 						</Button>
@@ -199,6 +227,7 @@ export default function MessageFeedback({
 								variant="ghost"
 								color="secondary"
 								onClick={onRegenerate}
+								aria-label="Regenerate response"
 							>
 								<RefreshCw size={12} />
 							</Button>

frontend/src/container/AIAssistant/components/UserMessageActions/UserMessageActions.tsx

@@ -47,6 +47,7 @@ export default function UserMessageActions({
 					variant="ghost"
 					color="secondary"
 					onClick={handleCopy}
+					aria-label={copied ? 'Copied' : 'Copy message'}
 				>
 					{copied ? <Check size={12} /> : <Copy size={12} />}
 				</Button>

frontend/src/container/AIAssistant/components/VirtualizedMessages/VirtualizedMessages.tsx

@@ -90,6 +90,16 @@ export default function VirtualizedMessages({
 
 	const virtuosoRef = useRef<VirtuosoHandle>(null);
 	const scrollerRef = useRef<HTMLElement | Window | null>(null);
+	// Tracks whether the scroller is pinned to (or near) the bottom. Updated
+	// via Virtuoso's `atBottomStateChange` so we can stop force-scrolling the
+	// user back down when they've intentionally scrolled up to read earlier
+	// content.
+	const atBottomRef = useRef(true);
+	// Id of the latest user message we've already anchored to. Used to detect
+	// a fresh user send so we can re-anchor to the bottom regardless of where
+	// the user was scrolled — sending a message and not seeing it is worse
+	// than the anti-yank guarantee.
+	const lastSeenUserMessageIdRef = useRef<string | null>(null);
 
 	const handleRegenerate = useCallback(
 		(messageId: string): void => {
@@ -111,8 +121,25 @@ export default function VirtualizedMessages({
 	// align: 'end')` would only reach the last item's bottom and leave the
 	// padding hidden below the fold. Use `auto` while streaming so the bottom
 	// stays glued as text deltas arrive; `smooth` lags when triggered every
-	// few ms.
+	// few ms. Bail out if the user has scrolled away from the bottom — that's
+	// an explicit signal they want to read earlier content without being
+	// yanked back.
 	useEffect(() => {
+		const lastMessage = messages[messages.length - 1];
+		const isFreshUserSend =
+			lastMessage?.role === 'user' &&
+			lastMessage.id !== lastSeenUserMessageIdRef.current;
+		if (isFreshUserSend) {
+			lastSeenUserMessageIdRef.current = lastMessage.id;
+			// Re-anchor so the user sees their own send (and the assistant's
+			// follow-up streaming) even if they were reading history when they
+			// hit Enter.
+			atBottomRef.current = true;
+		}
+
+		if (!atBottomRef.current) {
+			return;
+		}
 		const scroller = scrollerRef.current;
 		if (!(scroller instanceof HTMLElement)) {
 			return;
@@ -122,7 +149,7 @@ export default function VirtualizedMessages({
 			behavior: isStreaming ? 'auto' : 'smooth',
 		});
 	}, [
-		messages.length,
+		messages,
 		streamingEvents.length,
 		streamingContentLength,
 		isStreaming,
@@ -132,14 +159,18 @@ export default function VirtualizedMessages({
 
 	const followOutput = useCallback(
 		(atBottom: boolean): false | 'auto' | 'smooth' => {
-			if (isStreaming) {
-				return 'auto';
+			if (!atBottom) {
+				return false;
 			}
-			return atBottom ? 'smooth' : false;
+			return isStreaming ? 'auto' : 'smooth';
 		},
 		[isStreaming],
 	);
 
+	const handleAtBottomStateChange = useCallback((atBottom: boolean): void => {
+		atBottomRef.current = atBottom;
+	}, []);
+
 	const showStreamingSlot =
 		isStreaming || Boolean(pendingApproval) || Boolean(pendingClarification);
 
@@ -188,6 +219,8 @@ export default function VirtualizedMessages({
 			className={styles.messages}
 			totalCount={totalCount}
 			followOutput={followOutput}
+			atBottomStateChange={handleAtBottomStateChange}
+			atBottomThreshold={64}
 			initialTopMostItemIndex={Math.max(0, totalCount - 1)}
 			itemContent={(index): JSX.Element => {
 				if (index < messages.length) {

frontend/src/container/AIAssistant/components/blocks/RichCodeBlock/RichCodeBlock.tsx

@@ -1,6 +1,7 @@
 import React, { useEffect, useRef, useState } from 'react';
 import { useCopyToClipboard } from 'react-use';
 import { Button } from '@signozhq/ui/button';
+import { TooltipSimple } from '@signozhq/ui/tooltip';
 import { Check, Copy } from '@signozhq/icons';
 import SyntaxHighlighter, {
 	a11yDark,
@@ -126,16 +127,17 @@ function CopyButton({ text }: { text: string }): JSX.Element {
 	};
 
 	return (
-		<Button
-			variant="ghost"
-			size="sm"
-			color="secondary"
-			className={styles.copyBtn}
-			onClick={handleCopy}
-			title={copied ? 'Copied' : 'Copy code'}
-			aria-label={copied ? 'Copied' : 'Copy code'}
-		>
-			{copied ? <Check size={12} /> : <Copy size={12} />}
-		</Button>
+		<TooltipSimple title={copied ? 'Copied' : 'Copy code'}>
+			<Button
+				variant="ghost"
+				size="sm"
+				color="secondary"
+				className={styles.copyBtn}
+				onClick={handleCopy}
+				aria-label={copied ? 'Copied' : 'Copy code'}
+			>
+				{copied ? <Check size={12} /> : <Copy size={12} />}
+			</Button>
+		</TooltipSimple>
 	);
 }

frontend/src/container/AIAssistant/events.ts

@@ -63,6 +63,26 @@ export const SuggestedPromptCategory = {
 export type SuggestedPromptCategory =
 	(typeof SuggestedPromptCategory)[keyof typeof SuggestedPromptCategory];
 
+// `source` attribute on the AI Assistant `Opened` event — describes which
+// surface triggered the open. Keep values stable: dashboards downstream
+// depend on the literal strings.
+export const AIAssistantOpenSource = {
+	Icon: 'icon',
+	Shortcut: 'shortcut',
+	Cmdk: 'cmdk',
+} as const;
+export type AIAssistantOpenSource =
+	(typeof AIAssistantOpenSource)[keyof typeof AIAssistantOpenSource];
+
+// `source` attribute on the `VoiceInputUsed` event — which surface initiated
+// the recording.
+export const VoiceInputSource = {
+	Button: 'button',
+	Shortcut: 'shortcut',
+} as const;
+export type VoiceInputSource =
+	(typeof VoiceInputSource)[keyof typeof VoiceInputSource];
+
 export enum AIAssistantEvents {
 	Opened = 'AI Assistant: Opened',
 	MessageSent = 'AI Assistant: Message sent',

frontend/src/container/AppLayout/utils.ts

@@ -1,9 +1,32 @@
 import ROUTES from 'constants/routes';
 
-export function getRouteKey(pathname: string): string {
-	const [routeKey] = Object.entries(ROUTES).find(
-		([, value]) => value === pathname,
-	) || ['DEFAULT'];
+const PARAM_SEGMENT = /:[^/]+/g;
+const REGEX_SPECIALS = /[.+*?^$()[\]{}|\\]/g;
 
-	return routeKey;
+function templateToRegex(template: string): RegExp {
+	const pattern = template
+		.replace(REGEX_SPECIALS, '\\$&')
+		.replace(PARAM_SEGMENT, '[^/]+');
+	return new RegExp(`^${pattern}$`);
+}
+
+export function getRouteKey(pathname: string): string {
+	const entries = Object.entries(ROUTES);
+
+	const exact = entries.find(([, value]) => value === pathname);
+	if (exact) {
+		return exact[0];
+	}
+
+	// First template that matches wins, so declaration order in `ROUTES`
+	// matters when templates can overlap. Today's set is unambiguous because
+	// `[^/]+` is segment-bounded, but if you ever add a sibling like
+	// `/services/list` next to `SERVICE_METRICS: '/services/:servicename'`,
+	// list the more-specific (more-static-segments) entry first in `ROUTES`
+	// — otherwise the param template will swallow the static path.
+	const dynamic = entries.find(
+		([, value]) => value.includes(':') && templateToRegex(value).test(pathname),
+	);
+
+	return dynamic?.[0] ?? 'DEFAULT';
 }

pkg/errors/errors.go

@@ -4,12 +4,12 @@ import (
 	"errors" //nolint:depguard
 	"fmt"
 	"log/slog"
+	"time"
 
 	"go.opentelemetry.io/otel/attribute"
 )
 
 // base is the fundamental struct that implements the error interface.
-// The order of the struct is 'TCMEUAS'.
 type base struct {
 	// t denotes the custom type of the error.
 	t typ
@@ -25,6 +25,12 @@ type base struct {
 	a []string
 	// s contains the stacktrace captured at error creation time.
 	s fmt.Stringer
+	// r is the retry strategy for the error, if applicable.
+	r *retry
+	// suggestions is a list of user-facing suggestions related to the error, if present.
+	suggestions []string
+	// invalidReferences is a list of references that were invalid and contributed to the error, if present.
+	invalidReferences []string
 }
 
 // Stacktrace returns the stacktrace captured at error creation time, formatted as a string.
@@ -39,13 +45,16 @@ func (b *base) Stacktrace() string {
 // and returns a new base error.
 func (b *base) WithStacktrace(s string) *base {
 	return &base{
-		t: b.t,
-		c: b.c,
-		m: b.m,
-		e: b.e,
-		u: b.u,
-		a: b.a,
-		s: rawStacktrace(s),
+		t:                 b.t,
+		c:                 b.c,
+		m:                 b.m,
+		e:                 b.e,
+		u:                 b.u,
+		a:                 b.a,
+		s:                 rawStacktrace(s),
+		r:                 b.r,
+		suggestions:       b.suggestions,
+		invalidReferences: b.invalidReferences,
 	}
 }
 
@@ -113,13 +122,16 @@ func WithAdditionalf(cause error, format string, args ...any) *base {
 		s = original.s
 	}
 	b := &base{
-		t: t,
-		c: c,
-		m: m,
-		e: e,
-		u: u,
-		a: a,
-		s: s,
+		t:                 t,
+		c:                 c,
+		m:                 m,
+		e:                 e,
+		u:                 u,
+		a:                 a,
+		s:                 s,
+		r:                 retryOf(cause),
+		suggestions:       suggestionsOf(cause),
+		invalidReferences: invalidReferencesOf(cause),
 	}
 
 	return b.WithAdditional(append(a, fmt.Sprintf(format, args...))...)
@@ -128,29 +140,113 @@ func WithAdditionalf(cause error, format string, args ...any) *base {
 // WithUrl adds a url to the base error and returns a new base error.
 func (b *base) WithUrl(u string) *base {
 	return &base{
-		t: b.t,
-		c: b.c,
-		m: b.m,
-		e: b.e,
-		u: u,
-		a: b.a,
-		s: b.s,
+		t:                 b.t,
+		c:                 b.c,
+		m:                 b.m,
+		e:                 b.e,
+		u:                 u,
+		a:                 b.a,
+		s:                 b.s,
+		r:                 b.r,
+		suggestions:       b.suggestions,
+		invalidReferences: b.invalidReferences,
 	}
 }
 
 // WithAdditional adds additional messages to the base error and returns a new base error.
 func (b *base) WithAdditional(a ...string) *base {
 	return &base{
-		t: b.t,
-		c: b.c,
-		m: b.m,
-		e: b.e,
-		u: b.u,
-		a: a,
-		s: b.s,
+		t:                 b.t,
+		c:                 b.c,
+		m:                 b.m,
+		e:                 b.e,
+		u:                 b.u,
+		a:                 a,
+		s:                 b.s,
+		r:                 b.r,
+		suggestions:       b.suggestions,
+		invalidReferences: b.invalidReferences,
 	}
 }
 
+// withRetry adds retry metadata to the base error and returns a new base error.
+func (b *base) withRetry(r retry) *base {
+	return &base{
+		t:                 b.t,
+		c:                 b.c,
+		m:                 b.m,
+		e:                 b.e,
+		u:                 b.u,
+		a:                 b.a,
+		s:                 b.s,
+		r:                 &r,
+		suggestions:       b.suggestions,
+		invalidReferences: b.invalidReferences,
+	}
+}
+
+// WithSuggestions replaces the list of suggestions on the base error.
+func (b *base) WithSuggestions(suggestions ...string) *base {
+	return &base{
+		t:                 b.t,
+		c:                 b.c,
+		m:                 b.m,
+		e:                 b.e,
+		u:                 b.u,
+		a:                 b.a,
+		s:                 b.s,
+		r:                 b.r,
+		suggestions:       suggestions,
+		invalidReferences: b.invalidReferences,
+	}
+}
+
+// WithInvalidReferences replaces the list of invalid references on the base error.
+func (b *base) WithInvalidReferences(invalidReferences ...string) *base {
+	return &base{
+		t:                 b.t,
+		c:                 b.c,
+		m:                 b.m,
+		e:                 b.e,
+		u:                 b.u,
+		a:                 b.a,
+		s:                 b.s,
+		r:                 b.r,
+		suggestions:       b.suggestions,
+		invalidReferences: invalidReferences,
+	}
+}
+
+// WithRetryNever sets the retry policy to Never.
+func (b *base) WithRetryNever() *base {
+	return b.withRetry(retry{policy: RetryNever})
+}
+
+// WithRetryImmediate sets the retry policy to Immediate.
+func (b *base) WithRetryImmediate() *base {
+	return b.withRetry(retry{policy: RetryImmediate})
+}
+
+// WithRetryBackoff sets the retry policy to Backoff.
+func (b *base) WithRetryBackoff() *base {
+	return b.withRetry(retry{policy: RetryBackoff})
+}
+
+// WithRetryAfter sets the retry policy to After and requires a delay.
+func (b *base) WithRetryAfter(delay time.Duration) *base {
+	return b.withRetry(newRetryAfter(delay))
+}
+
+// WithRetryAfterFix sets the retry policy to AfterFix.
+func (b *base) WithRetryAfterFix() *base {
+	return b.withRetry(retry{policy: RetryAfterFix})
+}
+
+// WithRetryAfterAuth sets the retry policy to AfterAuth.
+func (b *base) WithRetryAfterAuth() *base {
+	return b.withRetry(retry{policy: RetryAfterAuth})
+}
+
 // Unwrapb is a combination of built-in errors.As and type casting.
 // It finds the first error in cause that matches base,
 // and if one is found, returns the individual fields of base.
@@ -198,57 +294,67 @@ func Is(err error, target error) bool {
 
 // WrapNotFoundf is a wrapper around Wrapf with TypeNotFound.
 func WrapNotFoundf(cause error, code Code, format string, args ...any) *base {
-	return Wrapf(cause, TypeNotFound, code, format, args...)
+	return Wrapf(cause, TypeNotFound, code, format, args...).withRetry(retry{policy: RetryNever})
 }
 
 // NewNotFoundf is a wrapper around Newf with TypeNotFound.
 func NewNotFoundf(code Code, format string, args ...any) *base {
-	return Newf(TypeNotFound, code, format, args...)
+	return Newf(TypeNotFound, code, format, args...).withRetry(retry{policy: RetryNever})
 }
 
 // WrapInternalf is a wrapper around Wrapf with TypeInternal.
 func WrapInternalf(cause error, code Code, format string, args ...any) *base {
-	return Wrapf(cause, TypeInternal, code, format, args...)
+	return Wrapf(cause, TypeInternal, code, format, args...).withRetry(retry{policy: RetryNever})
 }
 
 // NewInternalf is a wrapper around Newf with TypeInternal.
 func NewInternalf(code Code, format string, args ...any) *base {
-	return Newf(TypeInternal, code, format, args...)
+	return Newf(TypeInternal, code, format, args...).withRetry(retry{policy: RetryNever})
 }
 
 // WrapInvalidInputf is a wrapper around Wrapf with TypeInvalidInput.
 func WrapInvalidInputf(cause error, code Code, format string, args ...any) *base {
-	return Wrapf(cause, TypeInvalidInput, code, format, args...)
+	return Wrapf(cause, TypeInvalidInput, code, format, args...).withRetry(retry{policy: RetryAfterFix})
 }
 
 // NewInvalidInputf is a wrapper around Newf with TypeInvalidInput.
 func NewInvalidInputf(code Code, format string, args ...any) *base {
-	return Newf(TypeInvalidInput, code, format, args...)
-}
-
-// WrapUnexpectedf is a wrapper around Wrapf with TypeUnexpected.
-func WrapUnexpectedf(cause error, code Code, format string, args ...any) *base {
-	return Wrapf(cause, TypeInvalidInput, code, format, args...)
-}
-
-// NewUnexpectedf is a wrapper around Newf with TypeUnexpected.
-func NewUnexpectedf(code Code, format string, args ...any) *base {
-	return Newf(TypeInvalidInput, code, format, args...)
+	return Newf(TypeInvalidInput, code, format, args...).withRetry(retry{policy: RetryAfterFix})
 }
 
 // NewMethodNotAllowedf is a wrapper around Newf with TypeMethodNotAllowed.
 func NewMethodNotAllowedf(code Code, format string, args ...any) *base {
-	return Newf(TypeMethodNotAllowed, code, format, args...)
+	return Newf(TypeMethodNotAllowed, code, format, args...).withRetry(retry{policy: RetryNever})
 }
 
 // WrapTimeoutf is a wrapper around Wrapf with TypeTimeout.
 func WrapTimeoutf(cause error, code Code, format string, args ...any) *base {
-	return Wrapf(cause, TypeTimeout, code, format, args...)
+	return Wrapf(cause, TypeTimeout, code, format, args...).withRetry(retry{policy: RetryBackoff})
 }
 
 // NewTimeoutf is a wrapper around Newf with TypeTimeout.
 func NewTimeoutf(code Code, format string, args ...any) *base {
-	return Newf(TypeTimeout, code, format, args...)
+	return Newf(TypeTimeout, code, format, args...).withRetry(retry{policy: RetryBackoff})
+}
+
+// WrapUnauthenticatedf is a wrapper around Wrapf with TypeUnauthenticated.
+func WrapUnauthenticatedf(cause error, code Code, format string, args ...any) *base {
+	return Wrapf(cause, TypeUnauthenticated, code, format, args...).withRetry(retry{policy: RetryAfterAuth})
+}
+
+// NewUnauthenticatedf is a wrapper around Newf with TypeUnauthenticated.
+func NewUnauthenticatedf(code Code, format string, args ...any) *base {
+	return Newf(TypeUnauthenticated, code, format, args...).withRetry(retry{policy: RetryAfterAuth})
+}
+
+// WrapForbiddenf is a wrapper around Wrapf with TypeForbidden.
+func WrapForbiddenf(cause error, code Code, format string, args ...any) *base {
+	return Wrapf(cause, TypeForbidden, code, format, args...).withRetry(retry{policy: RetryNever})
+}
+
+// NewForbiddenf is a wrapper around Newf with TypeForbidden.
+func NewForbiddenf(code Code, format string, args ...any) *base {
+	return Newf(TypeForbidden, code, format, args...).withRetry(retry{policy: RetryNever})
 }
 
 // Attr returns an slog.Attr with a standardized "exception" key for the given error.
@@ -262,3 +368,36 @@ func TypeAttr(err error) attribute.KeyValue {
 	t, _, _, _, _, _ := Unwrapb(err)
 	return attribute.String("error.type", t.String())
 }
+
+// RetryAfterOf returns the explicit retry delay
+func RetryDelayOf(err error) time.Duration {
+	base, ok := err.(*base)
+	if !ok || base.r == nil || base.r.policy != RetryAfter {
+		return 0
+	}
+	return base.r.delay
+}
+
+func retryOf(err error) *retry {
+	base, ok := err.(*base)
+	if ok {
+		return base.r
+	}
+	return nil
+}
+
+func suggestionsOf(err error) []string {
+	base, ok := err.(*base)
+	if ok {
+		return base.suggestions
+	}
+	return nil
+}
+
+func invalidReferencesOf(err error) []string {
+	base, ok := err.(*base)
+	if ok {
+		return base.invalidReferences
+	}
+	return nil
+}

pkg/errors/errors_test.go

@@ -3,8 +3,10 @@ package errors
 import (
 	"errors" //nolint:depguard
 	"testing"
+	"time"
 
 	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
 )
 
 func TestNew(t *testing.T) {
@@ -59,6 +61,124 @@ func TestAttr(t *testing.T) {
 	assert.Equal(t, err, attr.Value.Any())
 }
 
+func TestWithSuggestions(t *testing.T) {
+	err := New(TypeInternal, MustNewCode("test_code"), "test error").WithSuggestions("try this")
+	assert.Equal(t, []string{"try this"}, suggestionsOf(err))
+
+	// WithSuggestions replaces the existing list.
+	err = err.WithSuggestions("try this instead")
+	assert.Equal(t, []string{"try this instead"}, suggestionsOf(err))
+
+	// Variadic form replaces with multiple entries.
+	err = err.WithSuggestions("first", "second")
+	assert.Equal(t, []string{"first", "second"}, suggestionsOf(err))
+}
+
+func TestWithRetryNever(t *testing.T) {
+	err := New(TypeInternal, MustNewCode("test_code"), "test error").WithRetryNever()
+	assert.Equal(t, RetryNever, retryOf(err).policy)
+}
+
+func TestWithRetryImmediate(t *testing.T) {
+	err := New(TypeInternal, MustNewCode("test_code"), "test error").WithRetryImmediate()
+	assert.Equal(t, RetryImmediate, retryOf(err).policy)
+}
+
+func TestWithRetryBackoff(t *testing.T) {
+	err := New(TypeInternal, MustNewCode("test_code"), "test error").WithRetryBackoff()
+	assert.Equal(t, RetryBackoff, retryOf(err).policy)
+}
+
+func TestWithRetryAfter(t *testing.T) {
+	err := New(TypeInternal, MustNewCode("test_code"), "test error").WithRetryAfter(5 * time.Microsecond)
+	r := retryOf(err)
+
+	assert.Equal(t, RetryAfter, r.policy)
+	assert.Equal(t, 5, int(r.delay.Microseconds()))
+}
+
+func TestWithRetryAfterFix(t *testing.T) {
+	err := New(TypeInternal, MustNewCode("test_code"), "test error").WithRetryAfterFix()
+	assert.Equal(t, RetryAfterFix, retryOf(err).policy)
+}
+
+func TestWithRetryAfterAuth(t *testing.T) {
+	err := New(TypeInternal, MustNewCode("test_code"), "test error").WithRetryAfterAuth()
+	assert.Equal(t, RetryAfterAuth, retryOf(err).policy)
+}
+
+func TestWithInvalidReferences(t *testing.T) {
+	// WithInvalidReferences populates the list.
+	err := New(TypeInvalidInput, MustNewCode("bad_ref"), "bad ref").
+		WithInvalidReferences("queries[0]", "queries[1]")
+	assert.Equal(t, []string{"queries[0]", "queries[1]"}, invalidReferencesOf(err))
+
+	// WithInvalidReferences replaces the entire list on each call.
+	err = err.WithInvalidReferences("queries[2]")
+	assert.Equal(t, []string{"queries[2]"}, invalidReferencesOf(err),
+		"WithInvalidReferences must replace the entire list")
+}
+
+func TestAsJSONBaseError(t *testing.T) {
+	err := New(TypeInvalidInput, MustNewCode("bad_input"), "field foo is bad").
+		WithUrl("https://docs/bad_input").
+		WithAdditional("hint1", "hint2").
+		WithSuggestions("try this").
+		WithInvalidReferences("queries[0]")
+
+	j := AsJSON(err)
+
+	assert.Equal(t, "invalid-input", j.Type)
+	assert.Equal(t, "bad_input", j.Code)
+	assert.Equal(t, "field foo is bad", j.Message)
+	assert.Equal(t, "https://docs/bad_input", j.Url)
+	assert.Equal(t, []responseerroradditional{{Message: "hint1"}, {Message: "hint2"}}, j.Errors)
+
+	// InvalidInput auto-applies the after_fix policy via NewInvalidInputf — but
+	// New (bare constructor) does not. The retry block should reflect that.
+	assert.Nil(t, j.Retry, "bare New(...) should not populate a retry block")
+
+	assert.Equal(t, []string{"try this"}, j.Suggestions)
+	assert.Equal(t, []string{"queries[0]"}, j.InvalidReferences)
+}
+
+func TestAsJSONRetryBlock(t *testing.T) {
+	t.Run("RetryAfterIncludesDuration", func(t *testing.T) {
+		err := NewTimeoutf(MustNewCode("slow"), "slow").WithRetryAfter(5 * time.Second)
+		j := AsJSON(err)
+		require.NotNil(t, j.Retry)
+		assert.Equal(t, responseretrypolicy(RetryAfter), j.Retry.Policy)
+		assert.Equal(t, "5s", j.Retry.Delay)
+	})
+
+	t.Run("NonAfterPolicyOmitsDurationField", func(t *testing.T) {
+		// NewInvalidInputf auto-applies retryAfterFix via the constructor helper.
+		err := NewInvalidInputf(MustNewCode("bad"), "bad")
+		j := AsJSON(err)
+		require.NotNil(t, j.Retry)
+		assert.Equal(t, responseretrypolicy(RetryAfterFix), j.Retry.Policy)
+		assert.Empty(t, j.Retry.Delay, "delay must be empty when policy != after")
+	})
+
+	t.Run("BareErrorOmitsRetryBlock", func(t *testing.T) {
+		err := New(TypeInternal, MustNewCode("boom"), "boom")
+		j := AsJSON(err)
+		assert.Nil(t, j.Retry, "bare New(...) without WithRetry* must omit retry")
+	})
+
+	t.Run("NonBaseErrorOmitsRetryBlock", func(t *testing.T) {
+		// Stdlib errors carry no retry metadata; AsJSON omits the retry block.
+		j := AsJSON(errors.New("plain stdlib error"))
+		assert.Nil(t, j.Retry, "non-base errors must omit the retry block")
+	})
+}
+
+func TestAsJSONOptionalFieldsOmittedWhenEmpty(t *testing.T) {
+	j := AsJSON(New(TypeInternal, MustNewCode("boom"), "boom"))
+	assert.Nil(t, j.Suggestions, "no suggestions set => Suggestions must be nil so json omitempty drops it")
+	assert.Nil(t, j.InvalidReferences, "no invalid references set => InvalidReferences must be nil so json omitempty drops it")
+}
+
 func TestWithStacktrace(t *testing.T) {
 	err := New(TypeInternal, MustNewCode("test_code"), "panic").WithStacktrace("custom stack trace")
 

pkg/errors/http.go

@@ -3,13 +3,38 @@ package errors
 import (
 	"encoding/json"
 	"net/url"
+	"time"
 )
 
 type JSON struct {
-	Code    string                    `json:"code" required:"true"`
-	Message string                    `json:"message" required:"true"`
-	Url     string                    `json:"url,omitempty"`
-	Errors  []responseerroradditional `json:"errors,omitempty"`
+	Type              string                    `json:"type,omitempty"`
+	Code              string                    `json:"code" required:"true"`
+	Message           string                    `json:"message" required:"true"`
+	Url               string                    `json:"url,omitempty"`
+	Errors            []responseerroradditional `json:"errors,omitempty"`
+	Retry             *responseretryjson        `json:"retry,omitempty"`
+	Suggestions       []string                  `json:"suggestions,omitempty"`
+	InvalidReferences []string                  `json:"invalidReferences,omitempty"`
+}
+
+type responseretryjson struct {
+	Policy responseretrypolicy `json:"policy"`
+	Delay  time.Duration       `json:"delay,omitempty"`
+}
+
+type responseretrypolicy string
+
+func (r responseretrypolicy) String() string { return string(r) }
+
+func (responseretrypolicy) Enum() []any {
+	return []any{
+		RetryNever,
+		RetryImmediate,
+		RetryBackoff,
+		RetryAfter,
+		RetryAfterFix,
+		RetryAfterAuth,
+	}
 }
 
 type responseerroradditional struct {
@@ -18,18 +43,30 @@ type responseerroradditional struct {
 
 func AsJSON(cause error) *JSON {
 	// See if this is an instance of the base error or not
-	_, c, m, _, u, a := Unwrapb(cause)
+	t, c, m, _, u, a := Unwrapb(cause)
 
 	rea := make([]responseerroradditional, len(a))
 	for k, v := range a {
 		rea[k] = responseerroradditional{v}
 	}
 
+	var retry *responseretryjson
+	if r := retryOf(cause); r != nil {
+		retry = &responseretryjson{Policy: responseretrypolicy(r.policy)}
+		if r.policy == RetryAfter {
+			retry.Delay = r.delay
+		}
+	}
+
 	return &JSON{
-		Code:    c.String(),
-		Message: m,
-		Url:     u,
-		Errors:  rea,
+		Type:              t.String(),
+		Code:              c.String(),
+		Message:           m,
+		Url:               u,
+		Errors:            rea,
+		Retry:             retry,
+		Suggestions:       suggestionsOf(cause),
+		InvalidReferences: invalidReferencesOf(cause),
 	}
 }
 

pkg/errors/retry.go

@@ -0,0 +1,31 @@
+package errors
+
+import (
+	"time"
+)
+
+type RetryPolicy string
+
+const (
+	RetryNever     RetryPolicy = "never"      // retry with the same inputs cannot succeed.
+	RetryImmediate RetryPolicy = "immediate"  // retry without waiting.
+	RetryBackoff   RetryPolicy = "backoff"    // caller picks its own backoff schedule.
+	RetryAfter     RetryPolicy = "after"      // honor Retry.After exactly (the producer knows the wait).
+	RetryAfterFix  RetryPolicy = "after_fix"  // retry pointless until the caller fixes the request.
+	RetryAfterAuth RetryPolicy = "after_auth" // retry pointless until the caller re-authenticates.
+)
+
+// retry pairs a RetryPolicy with the canonical gRPC RetryInfo detail.
+// info is non-nil only when policy == RetryAfter.
+type retry struct {
+	policy RetryPolicy
+	delay  time.Duration
+}
+
+// newRetryAfter builds a retry value carrying a gRPC RetryInfo with the given delay.
+func newRetryAfter(d time.Duration) retry {
+	return retry{
+		policy: RetryAfter,
+		delay:  d,
+	}
+}

pkg/errors/type.go

@@ -11,8 +11,7 @@ var (
 	TypeForbidden              = typ{"forbidden"}
 	TypeCanceled               = typ{"canceled"}
 	TypeTimeout                = typ{"timeout"}
-	TypeUnexpected             = typ{"unexpected"} // Generic mismatch of expectations
-	TypeFatal                  = typ{"fatal"}      // Unrecoverable failure (e.g. panic)
+	TypeFatal                  = typ{"fatal"} // Unrecoverable failure (e.g. panic)
 	TypeLicenseUnavailable     = typ{"license-unavailable"}
 	TypeTooManyRequests        = typ{"too-many-requests"}
 )

pkg/http/render/render.go

@@ -1,7 +1,9 @@
 package render
 
 import (
+	"math"
 	"net/http"
+	"strconv"
 
 	"github.com/SigNoz/signoz/pkg/errors"
 	jsoniter "github.com/json-iterator/go"
@@ -121,6 +123,14 @@ func Error(rw http.ResponseWriter, cause error) {
 		return
 	}
 
+	// Retry-After carries the explicit delay declared via
+	// errors.WithRetryAfter. Set it before WriteHeader so headers go on the wire.
+	d := errors.RetryDelayOf(cause)
+	if d.Seconds() > 0 {
+		rw.Header().Set("Retry-After", strconv.Itoa(int(math.Ceil(d.Seconds()))))
+
+	}
+
 	rw.WriteHeader(httpCode)
 	_, _ = rw.Write(body)
 }

pkg/http/render/render_test.go

@@ -6,7 +6,9 @@ import (
 	"io"
 	"net"
 	"net/http"
+	"net/http/httptest"
 	"testing"
+	"time"
 
 	"github.com/SigNoz/signoz/pkg/errors"
 	"github.com/stretchr/testify/assert"
@@ -97,13 +99,13 @@ func TestError(t *testing.T) {
 			name:       "AlreadyExists",
 			statusCode: http.StatusConflict,
 			err:        errors.New(errors.TypeAlreadyExists, errors.MustNewCode("already_exists"), "already exists").WithUrl("https://already_exists"),
-			expected:   []byte(`{"status":"error","error":{"code":"already_exists","message":"already exists","url":"https://already_exists"}}`),
+			expected:   []byte(`{"status":"error","error":{"type":"already-exists","code":"already_exists","message":"already exists","url":"https://already_exists"}}`),
 		},
 		"/unauthenticated": {
 			name:       "Unauthenticated",
 			statusCode: http.StatusUnauthorized,
 			err:        errors.New(errors.TypeUnauthenticated, errors.MustNewCode("not_allowed"), "not allowed").WithUrl("https://unauthenticated").WithAdditional("a1", "a2"),
-			expected:   []byte(`{"status":"error","error":{"code":"not_allowed","message":"not allowed","url":"https://unauthenticated","errors":[{"message":"a1"},{"message":"a2"}]}}`),
+			expected:   []byte(`{"status":"error","error":{"type":"unauthenticated","code":"not_allowed","message":"not allowed","url":"https://unauthenticated","errors":[{"message":"a1"},{"message":"a2"}]}}`),
 		},
 	}
 
@@ -145,3 +147,79 @@ func TestError(t *testing.T) {
 	}
 
 }
+
+// TestErrorRetryAfterHeader verifies that the HTTP Retry-After header is set
+// when (and only when) the error declares an explicit WithRetryAfter delay.
+// Other retry policies (backoff, after_fix, after_auth, never) and bare errors
+// without any policy must NOT emit the header — clients ignore non-numeric or
+// missing values, but emitting one wrongly would mislead retry libraries.
+func TestErrorRetryAfterHeader(t *testing.T) {
+	testCases := map[string]struct {
+		name                string
+		err                 error
+		wantRetryAfter      string // expected header value; "" means header must be absent
+		wantBodyContains    string // substring that must appear in the JSON body
+		wantBodyNotContains string // substring that must NOT appear in the JSON body
+	}{
+		"/with_retry_after_5s": {
+			name:             "ExplicitDelay5Seconds",
+			err:              errors.New(errors.TypeTooManyRequests, errors.MustNewCode("rate_limited"), "slow down").WithRetryAfter(5 * time.Second),
+			wantRetryAfter:   "5",
+			wantBodyContains: `"retry":{"policy":"after","delay":5000000000}`,
+		},
+		"/with_retry_after_subsecond": {
+			name:             "SubSecondRoundsUp",
+			err:              errors.New(errors.TypeTooManyRequests, errors.MustNewCode("rate_limited"), "slow down").WithRetryAfter(500 * time.Millisecond),
+			wantRetryAfter:   "1", // ceiling-rounded
+			wantBodyContains: `"delay":500000000`,
+		},
+		"/timeout_uses_backoff": {
+			name:             "BackoffPolicyNoHeader",
+			err:              errors.NewTimeoutf(errors.MustNewCode("slow_query"), "query timed out"),
+			wantRetryAfter:   "",
+			wantBodyContains: `"retry":{"policy":"backoff"}`,
+		},
+		"/invalid_input_after_fix": {
+			name:             "AfterFixPolicyNoHeader",
+			err:              errors.NewInvalidInputf(errors.MustNewCode("bad_field"), "bad field"),
+			wantRetryAfter:   "",
+			wantBodyContains: `"retry":{"policy":"after_fix"}`,
+		},
+		"/bare_no_policy": {
+			name:                "BareErrorNoHeaderNoRetryBlock",
+			err:                 errors.New(errors.TypeInternal, errors.MustNewCode("boom"), "boom"),
+			wantRetryAfter:      "",
+			wantBodyContains:    `"code":"boom"`,
+			wantBodyNotContains: `"retry"`,
+		},
+	}
+
+	srv := httptest.NewServer(http.HandlerFunc(func(rw http.ResponseWriter, req *http.Request) {
+		if tc, ok := testCases[req.URL.Path]; ok {
+			Error(rw, tc.err)
+		}
+	}))
+	defer srv.Close()
+
+	for path, tc := range testCases {
+		t.Run(tc.name, func(t *testing.T) {
+			res, err := http.Get(srv.URL + path)
+			require.NoError(t, err)
+			defer func() { require.NoError(t, res.Body.Close()) }()
+
+			body, err := io.ReadAll(res.Body)
+			require.NoError(t, err)
+
+			assert.Equal(t, tc.wantRetryAfter, res.Header.Get("Retry-After"),
+				"Retry-After header for %s", tc.name)
+			if tc.wantBodyContains != "" {
+				assert.Contains(t, string(body), tc.wantBodyContains,
+					"body should contain %q for %s", tc.wantBodyContains, tc.name)
+			}
+			if tc.wantBodyNotContains != "" {
+				assert.NotContains(t, string(body), tc.wantBodyNotContains,
+					"body should NOT contain %q for %s", tc.wantBodyNotContains, tc.name)
+			}
+		})
+	}
+}

pkg/meterreporter/config.go

@@ -1,6 +1,7 @@
 package meterreporter
 
 import (
+	"math/rand/v2"
 	"time"
 
 	"github.com/SigNoz/signoz/pkg/errors"
@@ -15,12 +16,21 @@ type Config struct {
 
 	// Backfill enables sealed-day catch-up from the license creation day.
 	Backfill bool `mapstructure:"backfill"`
+
+	// Jitter is the randomness applied to both the first collect after
+	// Start() and to every subsequent cycle. The first fire happens at a
+	// random time in [0, Jitter); each subsequent cycle takes
+	// Interval - random(0, Jitter). Negative (the default) means "derive
+	// from Interval" via ResolvedJitter, so the value scales with whatever
+	// Interval the user picks.
+	Jitter time.Duration `mapstructure:"jitter"`
 }
 
 func newConfig() factory.Config {
 	return Config{
 		Interval: 6 * time.Hour,
 		Backfill: true,
+		Jitter:   -1, // Negative sentinel. Resolved at use time unless explicitly set.
 	}
 }
 
@@ -29,9 +39,27 @@ func NewConfigFactory() factory.ConfigFactory {
 }
 
 func (c Config) Validate() error {
-	if c.Interval < 5*time.Minute || c.Interval > 24*time.Hour {
-		return errors.New(errors.TypeInvalidInput, ErrCodeInvalidInput, "meterreporter::interval must be between 5m and 24h")
+	if c.Interval < 10*time.Minute || c.Interval > 24*time.Hour {
+		return errors.New(errors.TypeInvalidInput, ErrCodeInvalidInput, "meterreporter::interval must be between 10m and 24h")
+	}
+
+	if c.Jitter >= 0 && (c.Jitter < 10*time.Minute || c.Jitter > c.Interval) {
+		return errors.New(errors.TypeInvalidInput, ErrCodeInvalidInput, "meterreporter::jitter must be between 10m and interval")
 	}
 
 	return nil
 }
+
+// NewJitter returns a fresh random duration sampled uniformly from
+// [0, jitter), where jitter is the configured Jitter or, if the sentinel
+// default is still in place, min(Interval, 2h).
+func (c Config) NewJitter() time.Duration {
+	defaultJitter := 2 * time.Hour
+
+	cap := c.Jitter
+	if cap < 0 {
+		cap = min(c.Interval, defaultJitter)
+	}
+
+	return time.Duration(rand.Int64N(int64(cap)))
+}

pkg/modules/rawdataexport/implrawdataexport/handler.go

@@ -202,15 +202,15 @@ func (handler *handler) exportRawDataJSONL(rowChan <-chan *qbtypes.RawRow, errCh
 			}
 			jsonBytes, err := json.Marshal(row.Data)
 			if err != nil {
-				return false, errors.NewUnexpectedf(errors.CodeInternal, "error marshaling JSON: %s", err)
+				return false, errors.NewInternalf(errors.CodeInternal, "error marshaling JSON: %s", err)
 			}
 			totalBytes += uint64(len(jsonBytes)) + 1
 
 			if _, err := writer.Write(jsonBytes); err != nil {
-				return false, errors.NewUnexpectedf(errors.CodeInternal, "error writing JSON: %s", err)
+				return false, errors.NewInternalf(errors.CodeInternal, "error writing JSON: %s", err)
 			}
 			if _, err := writer.Write([]byte("\n")); err != nil {
-				return false, errors.NewUnexpectedf(errors.CodeInternal, "error writing JSON newline: %s", err)
+				return false, errors.NewInternalf(errors.CodeInternal, "error writing JSON newline: %s", err)
 			}
 
 			if totalBytes > MaxExportBytesLimit {

pkg/modules/user/impluser/getter.go

@@ -74,7 +74,7 @@ func (module *getter) ListDeprecatedUsersByOrgID(ctx context.Context, orgID valu
 		roleNames := userIDToRoleNames[user.ID]
 
 		if len(roleNames) == 0 {
-			return nil, errors.Newf(errors.TypeUnexpected, authtypes.ErrCodeUserRolesNotFound, "no user roles entries found for user: %s", user.ID.String())
+			return nil, errors.Newf(errors.TypeInternal, authtypes.ErrCodeUserRolesNotFound, "no user roles entries found for user: %s", user.ID.String())
 		}
 
 		role := authtypes.SigNozManagedRoleToExistingLegacyRole[roleNames[0]]
@@ -113,11 +113,11 @@ func (module *getter) GetDeprecatedUserByOrgIDAndID(ctx context.Context, orgID v
 	}
 
 	if len(userRoles) == 0 {
-		return nil, errors.New(errors.TypeUnexpected, authtypes.ErrCodeUserRolesNotFound, "no user roles entries found")
+		return nil, errors.New(errors.TypeInternal, authtypes.ErrCodeUserRolesNotFound, "no user roles entries found")
 	}
 
 	if userRoles[0].Role == nil {
-		return nil, errors.New(errors.TypeUnexpected, authtypes.ErrCodeRoleNotFound, "role not found for user role entry")
+		return nil, errors.New(errors.TypeInternal, authtypes.ErrCodeRoleNotFound, "role not found for user role entry")
 	}
 
 	role := authtypes.SigNozManagedRoleToExistingLegacyRole[userRoles[0].Role.Name]
@@ -141,11 +141,11 @@ func (module *getter) Get(ctx context.Context, id valuer.UUID) (*types.Deprecate
 	}
 
 	if len(userRoles) == 0 {
-		return nil, errors.New(errors.TypeUnexpected, authtypes.ErrCodeUserRolesNotFound, "no user roles entries found")
+		return nil, errors.New(errors.TypeInternal, authtypes.ErrCodeUserRolesNotFound, "no user roles entries found")
 	}
 
 	if userRoles[0].Role == nil {
-		return nil, errors.New(errors.TypeUnexpected, authtypes.ErrCodeRoleNotFound, "role not found for user role entry")
+		return nil, errors.New(errors.TypeInternal, authtypes.ErrCodeRoleNotFound, "role not found for user role entry")
 	}
 
 	role := authtypes.SigNozManagedRoleToExistingLegacyRole[userRoles[0].Role.Name]
@@ -211,7 +211,7 @@ func (module *getter) GetRolesByUserID(ctx context.Context, userID valuer.UUID)
 
 	for _, ur := range userRoles {
 		if ur.Role == nil {
-			return nil, errors.New(errors.TypeUnexpected, authtypes.ErrCodeRoleNotFound, "role not found for user role entry")
+			return nil, errors.New(errors.TypeInternal, authtypes.ErrCodeRoleNotFound, "role not found for user role entry")
 		}
 	}
 

pkg/query-service/app/logparsingpipeline/collector_config.go

@@ -2,11 +2,14 @@ package logparsingpipeline
 
 import (
 	"encoding/json"
+	"fmt"
 	"strings"
 	"sync"
 
 	"gopkg.in/yaml.v3"
 
+	"log/slog"
+
 	"github.com/SigNoz/signoz/pkg/errors"
 	"github.com/SigNoz/signoz/pkg/query-service/constants"
 	"github.com/SigNoz/signoz/pkg/types/pipelinetypes"
@@ -23,13 +26,6 @@ var (
 	CodeCollectorConfigLogsPipelineNotFound   = errors.MustNewCode("collector_config_logs_pipeline_not_found")
 )
 
-const (
-	memoryLimiterProcessor       = "memory_limiter"
-	memoryLimiterProcessorPrefix = "memory_limiter/"
-	batchProcessor               = "batch"
-	batchProcessorPrefix         = "batch/"
-)
-
 // check if the processors already exist
 // if yes then update the processor.
 // if something doesn't exists then remove it.
@@ -83,14 +79,6 @@ func getOtelPipelineFromConfig(config map[string]interface{}) (*otelPipeline, er
 	return &p, nil
 }
 
-// buildCollectorPipelineProcessorsList assembles the final processor list in the
-// required order:
-//
-//  1. memory_limiter processors (any processor named "memory_limiter" or "memory_limiter/<id>")
-//  2. other existing processors (in their original order), which may include signoz processors
-//     that are not user-pipeline processors
-//  3. signoz user-pipeline processors
-//  4. batch processors (any processor named "batch" or "batch/<id>")
 func buildCollectorPipelineProcessorsList(
 	currentCollectorProcessors []string,
 	signozPipelineProcessorNames []string,
@@ -98,59 +86,90 @@ func buildCollectorPipelineProcessorsList(
 	lockLogsPipelineSpec.Lock()
 	defer lockLogsPipelineSpec.Unlock()
 
-	// Build a set of the desired signoz processors so we can drop any stale version
-	// of them (regardless of how they got into the current config) without
-	// accidentally duplicating them in the output.
-	desiredUserPipelineSet := make(map[string]struct{}, len(signozPipelineProcessorNames))
-	for _, p := range signozPipelineProcessorNames {
-		desiredUserPipelineSet[p] = struct{}{}
+	exists := map[string]struct{}{}
+	for _, v := range signozPipelineProcessorNames {
+		exists[v] = struct{}{}
 	}
 
-	result := make([]string, 0, len(currentCollectorProcessors)+len(signozPipelineProcessorNames))
+	// removed the old processors which are not used
+	var pipeline []string
+	for _, procName := range currentCollectorProcessors {
+		_, isInDesiredPipelineProcs := exists[procName]
+		if isInDesiredPipelineProcs || !hasSignozPipelineProcessorPrefix(procName) {
+			pipeline = append(pipeline, procName)
+		}
+	}
 
-	// Note: logic assumes there'll be only one batch processor
-	var batchProcIdx int
-	var batchProcFound bool
-iteration:
-	for idx, p := range currentCollectorProcessors {
-		_, inDesiredSet := desiredUserPipelineSet[p]
-		switch {
-		// same processor exist; retain the location of pre-existing location
-		case p == memoryLimiterProcessor || strings.HasPrefix(p, memoryLimiterProcessorPrefix):
-			result = append(result, p)
-		case hasSignozPipelineProcessorPrefix(p):
-			// this processor has been dropped
-			if !inDesiredSet {
-				continue iteration
-			} else {
-				result = append(result, p)
+	// create a reverse map of existing config processors and their position
+	existing := map[string]int{}
+	for i, p := range pipeline {
+		name := p
+		existing[name] = i
+	}
+
+	// create mapping from our logsParserPipeline to position in existing processors (from current config)
+	// this means, if "batch" holds position 3 in the current effective config, and 2 in our config, the map will be [2]: 3
+	specVsExistingMap := map[int]int{}
+	existingVsSpec := map[int]int{}
+
+	// go through plan and map its elements to current positions in effective config
+	for i, m := range signozPipelineProcessorNames {
+		if loc, ok := existing[m]; ok {
+			specVsExistingMap[i] = loc
+			existingVsSpec[loc] = i
+		}
+	}
+
+	lastMatched := 0
+	newPipeline := []string{}
+
+	for i := 0; i < len(signozPipelineProcessorNames); i++ {
+		m := signozPipelineProcessorNames[i]
+		if loc, ok := specVsExistingMap[i]; ok {
+			for j := lastMatched; j < loc; j++ {
+				if hasSignozPipelineProcessorPrefix(pipeline[j]) {
+					delete(specVsExistingMap, existingVsSpec[j])
+				} else {
+					newPipeline = append(newPipeline, pipeline[j])
+				}
 			}
-		case p == batchProcessor || strings.HasPrefix(p, batchProcessorPrefix):
-			batchProcIdx = idx
-			batchProcFound = true
-			break iteration
-		default:
-			result = append(result, p)
+			newPipeline = append(newPipeline, pipeline[loc])
+			lastMatched = loc + 1
+		} else {
+			newPipeline = append(newPipeline, m)
 		}
 
-		if inDesiredSet {
-			// delete from desired pipeline set so they're not added twice
-			delete(desiredUserPipelineSet, p)
-		}
 	}
-	// add user pipelines
-	for _, proc := range signozPipelineProcessorNames {
-		_, add := desiredUserPipelineSet[proc]
-		if add {
-			result = append(result, proc)
-		}
+	if lastMatched < len(pipeline) {
+		newPipeline = append(newPipeline, pipeline[lastMatched:]...)
 	}
 
-	// add batch processor and rest
-	if batchProcFound {
-		result = append(result, currentCollectorProcessors[batchProcIdx:]...)
+	if checkDuplicateString(newPipeline) {
+		// duplicates are most likely because the processor sequence in effective config conflicts
+		// with the planned sequence as per planned pipeline
+		return pipeline, fmt.Errorf("the effective config has an unexpected processor sequence: %v", pipeline)
 	}
-	return result, nil
+
+	return newPipeline, nil
+}
+
+func checkDuplicateString(pipeline []string) bool {
+	exists := make(map[string]bool, len(pipeline))
+	slog.Debug("checking duplicate processors in the pipeline", "pipeline", pipeline)
+	for _, processor := range pipeline {
+		name := processor
+		if _, ok := exists[name]; ok {
+			slog.Error(
+				"duplicate processor name detected in generated collector config for log pipelines",
+				"processor", processor,
+				"pipeline", pipeline,
+			)
+			return true
+		}
+
+		exists[name] = true
+	}
+	return false
 }
 
 func GenerateCollectorConfigWithPipelines(config []byte, pipelines []pipelinetypes.GettablePipeline) ([]byte, error) {

pkg/query-service/app/logparsingpipeline/collector_config_test.go

@@ -106,109 +106,107 @@ func TestBuildLogParsingProcessors(t *testing.T) {
 }
 
 var BuildLogsPipelineTestData = []struct {
-	Name          string
-	fromCollector []string
-	userPipelines []string
-	finalOutput   []string
+	Name             string
+	currentPipeline  []string
+	logsPipeline     []string
+	expectedPipeline []string
 }{
 	{
-		Name:          "Add new pipelines",
-		fromCollector: []string{"processor1", "processor2"},
-		userPipelines: []string{constants.LogsPPLPfx + "a", constants.LogsPPLPfx + "b"},
-		finalOutput:   []string{"processor1", "processor2", constants.LogsPPLPfx + "a", constants.LogsPPLPfx + "b"},
+		Name:             "Add new pipelines",
+		currentPipeline:  []string{"processor1", "processor2"},
+		logsPipeline:     []string{constants.LogsPPLPfx + "a", constants.LogsPPLPfx + "b"},
+		expectedPipeline: []string{constants.LogsPPLPfx + "a", constants.LogsPPLPfx + "b", "processor1", "processor2"},
 	},
 	{
-		Name:          "Add new pipeline and respect custom processors",
-		fromCollector: []string{constants.LogsPPLPfx + "a", "processor1", constants.LogsPPLPfx + "b", "processor2"},
-		userPipelines: []string{constants.LogsPPLPfx + "a", constants.LogsPPLPfx + "b", constants.LogsPPLPfx + "c"},
-		finalOutput:   []string{constants.LogsPPLPfx + "a", "processor1", constants.LogsPPLPfx + "b", "processor2", constants.LogsPPLPfx + "c"},
+		Name:             "Add new pipeline and respect custom processors",
+		currentPipeline:  []string{constants.LogsPPLPfx + "a", "processor1", constants.LogsPPLPfx + "b", "processor2"},
+		logsPipeline:     []string{constants.LogsPPLPfx + "a", constants.LogsPPLPfx + "b", constants.LogsPPLPfx + "c"},
+		expectedPipeline: []string{constants.LogsPPLPfx + "a", "processor1", constants.LogsPPLPfx + "b", constants.LogsPPLPfx + "c", "processor2"},
 	},
 	{
-		Name:          "Add new pipeline and respect custom processors",
-		fromCollector: []string{constants.LogsPPLPfx + "a", "processor1", constants.LogsPPLPfx + "b", "processor2"},
-		userPipelines: []string{constants.LogsPPLPfx + "a", constants.LogsPPLPfx + "b", constants.LogsPPLPfx + "c", constants.LogsPPLPfx + "d"},
-		finalOutput:   []string{constants.LogsPPLPfx + "a", "processor1", constants.LogsPPLPfx + "b", "processor2", constants.LogsPPLPfx + "c", constants.LogsPPLPfx + "d"},
+		Name:             "Add new pipeline and respect custom processors",
+		currentPipeline:  []string{constants.LogsPPLPfx + "a", "processor1", constants.LogsPPLPfx + "b", "processor2"},
+		logsPipeline:     []string{constants.LogsPPLPfx + "a", constants.LogsPPLPfx + "b", constants.LogsPPLPfx + "c", constants.LogsPPLPfx + "d"},
+		expectedPipeline: []string{constants.LogsPPLPfx + "a", "processor1", constants.LogsPPLPfx + "b", constants.LogsPPLPfx + "c", constants.LogsPPLPfx + "d", "processor2"},
 	},
 	{
-		Name:          "Add new pipeline and respect custom processors in the beginning and middle",
-		fromCollector: []string{"processor1", constants.LogsPPLPfx + "a", "processor2", constants.LogsPPLPfx + "b", "batch"},
-		userPipelines: []string{constants.LogsPPLPfx + "a", constants.LogsPPLPfx + "b", constants.LogsPPLPfx + "c"},
-		finalOutput:   []string{"processor1", constants.LogsPPLPfx + "a", "processor2", constants.LogsPPLPfx + "b", constants.LogsPPLPfx + "c", "batch"},
+		Name:             "Add new pipeline and respect custom processors in the beginning and middle",
+		currentPipeline:  []string{"processor1", constants.LogsPPLPfx + "a", "processor2", constants.LogsPPLPfx + "b", "batch"},
+		logsPipeline:     []string{constants.LogsPPLPfx + "a", constants.LogsPPLPfx + "b", constants.LogsPPLPfx + "c"},
+		expectedPipeline: []string{"processor1", constants.LogsPPLPfx + "a", "processor2", constants.LogsPPLPfx + "b", constants.LogsPPLPfx + "c", "batch"},
 	},
 	{
-		Name:          "Remove old pipeline add add new",
-		fromCollector: []string{constants.LogsPPLPfx + "a", constants.LogsPPLPfx + "b", "processor1", "processor2"},
-		userPipelines: []string{constants.LogsPPLPfx + "a"},
-		finalOutput:   []string{constants.LogsPPLPfx + "a", "processor1", "processor2"},
+		Name:             "Remove old pipeline add add new",
+		currentPipeline:  []string{constants.LogsPPLPfx + "a", constants.LogsPPLPfx + "b", "processor1", "processor2"},
+		logsPipeline:     []string{constants.LogsPPLPfx + "a"},
+		expectedPipeline: []string{constants.LogsPPLPfx + "a", "processor1", "processor2"},
 	},
 	{
-		Name:          "Remove old pipeline from middle",
-		fromCollector: []string{"processor1", "processor2", constants.LogsPPLPfx + "a", "processor3", constants.LogsPPLPfx + "b", "batch"},
-		userPipelines: []string{constants.LogsPPLPfx + "a"},
-		finalOutput:   []string{"processor1", "processor2", constants.LogsPPLPfx + "a", "processor3", "batch"},
+		Name:             "Remove old pipeline from middle",
+		currentPipeline:  []string{"processor1", "processor2", constants.LogsPPLPfx + "a", "processor3", constants.LogsPPLPfx + "b", "batch"},
+		logsPipeline:     []string{constants.LogsPPLPfx + "a"},
+		expectedPipeline: []string{"processor1", "processor2", constants.LogsPPLPfx + "a", "processor3", "batch"},
 	},
 	{
-		Name:          "Remove old pipeline from middle and add new pipeline",
-		fromCollector: []string{"memory_limiter", "processor1", "processor2", constants.LogsPPLPfx + "a", "processor3", constants.LogsPPLPfx + "b", "batch"},
-		userPipelines: []string{constants.LogsPPLPfx + "a", constants.LogsPPLPfx + "c"},
-		finalOutput:   []string{"memory_limiter", "processor1", "processor2", constants.LogsPPLPfx + "a", "processor3", constants.LogsPPLPfx + "c", "batch"},
+		Name:             "Remove old pipeline from middle and add new pipeline",
+		currentPipeline:  []string{"processor1", "processor2", constants.LogsPPLPfx + "a", "processor3", constants.LogsPPLPfx + "b", "batch"},
+		logsPipeline:     []string{constants.LogsPPLPfx + "a", constants.LogsPPLPfx + "c"},
+		expectedPipeline: []string{"processor1", "processor2", constants.LogsPPLPfx + "a", constants.LogsPPLPfx + "c", "processor3", "batch"},
 	},
 	{
-		Name:          "Remove multiple old pipelines from middle and add multiple new ones",
-		fromCollector: []string{"processor1", constants.LogsPPLPfx + "a", "processor2", constants.LogsPPLPfx + "b", "processor3", constants.LogsPPLPfx + "c", "processor4", constants.LogsPPLPfx + "d", "processor5", "batch"},
-		userPipelines: []string{constants.LogsPPLPfx + "a", constants.LogsPPLPfx + "a1", constants.LogsPPLPfx + "c", constants.LogsPPLPfx + "c1"},
-		finalOutput:   []string{"processor1", constants.LogsPPLPfx + "a", "processor2", "processor3", constants.LogsPPLPfx + "c", "processor4", "processor5", constants.LogsPPLPfx + "a1", constants.LogsPPLPfx + "c1", "batch"},
+		Name:             "Remove multiple old pipelines from middle and add multiple new ones",
+		currentPipeline:  []string{"processor1", constants.LogsPPLPfx + "a", "processor2", constants.LogsPPLPfx + "b", "processor3", constants.LogsPPLPfx + "c", "processor4", constants.LogsPPLPfx + "d", "processor5", "batch"},
+		logsPipeline:     []string{constants.LogsPPLPfx + "a", constants.LogsPPLPfx + "a1", constants.LogsPPLPfx + "c", constants.LogsPPLPfx + "c1"},
+		expectedPipeline: []string{"processor1", constants.LogsPPLPfx + "a", constants.LogsPPLPfx + "a1", "processor2", "processor3", constants.LogsPPLPfx + "c", constants.LogsPPLPfx + "c1", "processor4", "processor5", "batch"},
+	},
+
+	// working
+	{
+		Name:             "rearrange pipelines",
+		currentPipeline:  []string{"processor1", "processor2", constants.LogsPPLPfx + "_a", "processor3", constants.LogsPPLPfx + "_b", "batch"},
+		logsPipeline:     []string{constants.LogsPPLPfx + "_b", constants.LogsPPLPfx + "_a"},
+		expectedPipeline: []string{"processor1", "processor2", "processor3", constants.LogsPPLPfx + "_b", constants.LogsPPLPfx + "_a", "batch"},
 	},
 	{
-		Name:          "rearrange pipelines",
-		fromCollector: []string{"processor1", "processor2", constants.LogsPPLPfx + "_a", "processor3", constants.LogsPPLPfx + "_b", "batch"},
-		userPipelines: []string{constants.LogsPPLPfx + "_b", constants.LogsPPLPfx + "_a"},
-		finalOutput:   []string{"processor1", "processor2", constants.LogsPPLPfx + "_a", "processor3", constants.LogsPPLPfx + "_b", "batch"},
+		Name:             "rearrange pipelines with new processor",
+		currentPipeline:  []string{"processor1", "processor2", constants.LogsPPLPfx + "_a", "processor3", constants.LogsPPLPfx + "_b", "batch"},
+		logsPipeline:     []string{constants.LogsPPLPfx + "_b", constants.LogsPPLPfx + "_a", constants.LogsPPLPfx + "_c"},
+		expectedPipeline: []string{"processor1", "processor2", "processor3", constants.LogsPPLPfx + "_b", constants.LogsPPLPfx + "_a", constants.LogsPPLPfx + "_c", "batch"},
+		// expectedPipeline: []string{"processor1", "processor2", constants.LogsPPLPfx + "_b", "processor3", constants.LogsPPLPfx + "_a", constants.LogsPPLPfx + "_c", "batch"},
 	},
 	{
-		Name:          "rearrange pipelines with new processor",
-		fromCollector: []string{"processor1", "processor2", constants.LogsPPLPfx + "_a", "processor3", constants.LogsPPLPfx + "_b", "batch"},
-		userPipelines: []string{constants.LogsPPLPfx + "_b", constants.LogsPPLPfx + "_a", constants.LogsPPLPfx + "_c"},
-		finalOutput:   []string{"processor1", "processor2", constants.LogsPPLPfx + "_a", "processor3", constants.LogsPPLPfx + "_b", constants.LogsPPLPfx + "_c", "batch"},
+		Name:             "delete processor",
+		currentPipeline:  []string{"processor1", "processor2", constants.LogsPPLPfx + "_a", "processor3", constants.LogsPPLPfx + "_b", "batch"},
+		logsPipeline:     []string{},
+		expectedPipeline: []string{"processor1", "processor2", "processor3", "batch"},
 	},
 	{
-		Name:          "delete processor",
-		fromCollector: []string{"processor1", "processor2", constants.LogsPPLPfx + "_a", "processor3", constants.LogsPPLPfx + "_b", "batch"},
-		userPipelines: []string{},
-		finalOutput:   []string{"processor1", "processor2", "processor3", "batch"},
+		Name:             "last to first",
+		currentPipeline:  []string{"processor1", "processor2", constants.LogsPPLPfx + "_a", "processor3", "processor4", constants.LogsPPLPfx + "_b", "batch", constants.LogsPPLPfx + "_c"},
+		logsPipeline:     []string{constants.LogsPPLPfx + "_c", constants.LogsPPLPfx + "_a", constants.LogsPPLPfx + "_b"},
+		expectedPipeline: []string{"processor1", "processor2", "processor3", "processor4", "batch", constants.LogsPPLPfx + "_c", constants.LogsPPLPfx + "_a", constants.LogsPPLPfx + "_b"},
 	},
 	{
-		Name:          "last to first",
-		fromCollector: []string{"processor1", "processor2", constants.LogsPPLPfx + "_a", "processor3", "processor4", constants.LogsPPLPfx + "_b", "batch", constants.LogsPPLPfx + "_c"},
-		userPipelines: []string{constants.LogsPPLPfx + "_c", constants.LogsPPLPfx + "_a", constants.LogsPPLPfx + "_b"},
-		finalOutput:   []string{"processor1", "processor2", constants.LogsPPLPfx + "_a", "processor3", "processor4", constants.LogsPPLPfx + "_b", constants.LogsPPLPfx + "_c", "batch", constants.LogsPPLPfx + "_c"},
+		Name:             "multiple rearrange pipelines",
+		currentPipeline:  []string{"processor1", "processor2", constants.LogsPPLPfx + "_a", "processor3", constants.LogsPPLPfx + "_b", "batch", constants.LogsPPLPfx + "_c", "processor4", "processor5", constants.LogsPPLPfx + "_d", "processor6", "processor7"},
+		logsPipeline:     []string{constants.LogsPPLPfx + "_b", constants.LogsPPLPfx + "_a", constants.LogsPPLPfx + "_d", constants.LogsPPLPfx + "_c", constants.LogsPPLPfx + "_e"},
+		expectedPipeline: []string{"processor1", "processor2", "processor3", constants.LogsPPLPfx + "_b", constants.LogsPPLPfx + "_a", "batch", "processor4", "processor5", constants.LogsPPLPfx + "_d", constants.LogsPPLPfx + "_c", constants.LogsPPLPfx + "_e", "processor6", "processor7"},
 	},
 	{
-		Name:          "multiple rearrange pipelines",
-		fromCollector: []string{"processor1", "processor2", constants.LogsPPLPfx + "_a", "processor3", constants.LogsPPLPfx + "_b", "batch", constants.LogsPPLPfx + "_c", "processor4", "processor5", constants.LogsPPLPfx + "_d", "processor6", "processor7"},
-		userPipelines: []string{constants.LogsPPLPfx + "_b", constants.LogsPPLPfx + "_a", constants.LogsPPLPfx + "_d", constants.LogsPPLPfx + "_c", constants.LogsPPLPfx + "_e"},
-		finalOutput:   []string{"processor1", "processor2", constants.LogsPPLPfx + "_a", "processor3", constants.LogsPPLPfx + "_b", constants.LogsPPLPfx + "_d", constants.LogsPPLPfx + "_c", constants.LogsPPLPfx + "_e", "batch", constants.LogsPPLPfx + "_c", "processor4", "processor5", constants.LogsPPLPfx + "_d", "processor6", "processor7"},
-	},
-	{
-		Name:          "multiple rearrange with new pipelines",
-		fromCollector: []string{"memory_limiter", "processor1", "processor2", constants.LogsPPLPfx + "_a", "processor3", constants.LogsPPLPfx + "_b", "batch", constants.LogsPPLPfx + "_c", "processor4", "processor5", constants.LogsPPLPfx + "_d", "processor6", "processor7"},
-		userPipelines: []string{constants.LogsPPLPfx + "_z", constants.LogsPPLPfx + "_b", constants.LogsPPLPfx + "_a", constants.LogsPPLPfx + "_d", constants.LogsPPLPfx + "_c", constants.LogsPPLPfx + "_e"},
-		finalOutput:   []string{"memory_limiter", "processor1", "processor2", constants.LogsPPLPfx + "_a", "processor3", constants.LogsPPLPfx + "_b", constants.LogsPPLPfx + "_z", constants.LogsPPLPfx + "_d", constants.LogsPPLPfx + "_c", constants.LogsPPLPfx + "_e", "batch", constants.LogsPPLPfx + "_c", "processor4", "processor5", constants.LogsPPLPfx + "_d", "processor6", "processor7"},
-	},
-	{
-		Name:          "Prefixed proc in desired set not duplicated from others",
-		fromCollector: []string{"memory_limiter/logs", "custom_proc", "resourcedetection", "batch/logs"},
-		userPipelines: []string{"custom_proc", constants.LogsPPLPfx + "a"},
-		finalOutput:   []string{"memory_limiter/logs", "custom_proc", "resourcedetection", constants.LogsPPLPfx + "a", "batch/logs"},
+		Name:             "multiple rearrange with new pipelines",
+		currentPipeline:  []string{"processor1", "processor2", constants.LogsPPLPfx + "_a", "processor3", constants.LogsPPLPfx + "_b", "batch", constants.LogsPPLPfx + "_c", "processor4", "processor5", constants.LogsPPLPfx + "_d", "processor6", "processor7"},
+		logsPipeline:     []string{constants.LogsPPLPfx + "_z", constants.LogsPPLPfx + "_b", constants.LogsPPLPfx + "_a", constants.LogsPPLPfx + "_d", constants.LogsPPLPfx + "_c", constants.LogsPPLPfx + "_e"},
+		expectedPipeline: []string{constants.LogsPPLPfx + "_z", "processor1", "processor2", "processor3", constants.LogsPPLPfx + "_b", constants.LogsPPLPfx + "_a", "batch", "processor4", "processor5", constants.LogsPPLPfx + "_d", constants.LogsPPLPfx + "_c", constants.LogsPPLPfx + "_e", "processor6", "processor7"},
 	},
 }
 
 func TestBuildLogsPipeline(t *testing.T) {
 	for _, test := range BuildLogsPipelineTestData {
 		Convey(test.Name, t, func() {
-			v, err := buildCollectorPipelineProcessorsList(test.fromCollector, test.userPipelines)
+			v, err := buildCollectorPipelineProcessorsList(test.currentPipeline, test.logsPipeline)
 			So(err, ShouldBeNil)
-			So(v, ShouldResemble, test.finalOutput)
+			fmt.Println(test.Name, "\n", test.currentPipeline, "\n", v, "\n", test.expectedPipeline)
+			So(v, ShouldResemble, test.expectedPipeline)
 		})
 	}
 }

pkg/query-service/app/logparsingpipeline/pipelineBuilder.go

@@ -216,7 +216,7 @@ func getOperators(ops []pipelinetypes.PipelineOperator) ([]pipelinetypes.Pipelin
 
 func processSeverityParser(operator *pipelinetypes.PipelineOperator) error {
 	if operator.Type != "severity_parser" {
-		return errors.NewUnexpectedf(CodeInvalidOperatorType, "operator type received %s", operator.Type)
+		return errors.NewInternalf(CodeInvalidOperatorType, "operator type received %s", operator.Type)
 	}
 
 	parseFromNotNilCheck, err := fieldNotNilCheck(operator.ParseFrom)
@@ -236,7 +236,7 @@ func processSeverityParser(operator *pipelinetypes.PipelineOperator) error {
 // processJSONParser converts simple JSON parser operator into multiple operators for JSONMapping of default variables
 func processJSONParser(parent *pipelinetypes.PipelineOperator) ([]pipelinetypes.PipelineOperator, error) {
 	if parent.Type != "json_parser" {
-		return nil, errors.NewUnexpectedf(CodeInvalidOperatorType, "operator type received %s", parent.Type)
+		return nil, errors.NewInternalf(CodeInvalidOperatorType, "operator type received %s", parent.Type)
 	}
 
 	parseFromNotNilCheck, err := fieldNotNilCheck(parent.ParseFrom)

pkg/signoz/provider.go

@@ -208,6 +208,7 @@ func NewSQLMigrationProviderFactories(
 		sqlmigration.NewMigrateCloudIntegrationDashboardsFactory(sqlstore),
 		sqlmigration.NewAddScopeToPlannedMaintenanceFactory(sqlstore, sqlschema),
 		sqlmigration.NewMigrateInstalledIntegrationDashboardsFactory(sqlstore),
+		sqlmigration.NewAddDashboardNameFactory(sqlstore, sqlschema),
 	)
 }
 

pkg/sqlmigration/089_add_dashboard_name.go

@@ -0,0 +1,85 @@
+package sqlmigration
+
+import (
+	"context"
+
+	"github.com/SigNoz/signoz/pkg/factory"
+	"github.com/SigNoz/signoz/pkg/sqlschema"
+	"github.com/SigNoz/signoz/pkg/sqlstore"
+	"github.com/uptrace/bun"
+	"github.com/uptrace/bun/migrate"
+)
+
+type addDashboardName struct {
+	sqlstore  sqlstore.SQLStore
+	sqlschema sqlschema.SQLSchema
+}
+
+func NewAddDashboardNameFactory(sqlstore sqlstore.SQLStore, sqlschema sqlschema.SQLSchema) factory.ProviderFactory[SQLMigration, Config] {
+	return factory.NewProviderFactory(
+		factory.MustNewName("add_dashboard_name"),
+		func(ctx context.Context, ps factory.ProviderSettings, c Config) (SQLMigration, error) {
+			return &addDashboardName{sqlstore: sqlstore, sqlschema: sqlschema}, nil
+		},
+	)
+}
+
+func (migration *addDashboardName) Register(migrations *migrate.Migrations) error {
+	return migrations.Register(migration.Up, migration.Down)
+}
+
+func (migration *addDashboardName) Up(ctx context.Context, db *bun.DB) error {
+	// dashboard is referenced by public_dashboard and integration_dashboard;
+	// FK enforcement must be off for the SQLite recreate-table fallback.
+	if err := migration.sqlschema.ToggleFKEnforcement(ctx, db, false); err != nil {
+		return err
+	}
+
+	tx, err := db.BeginTx(ctx, nil)
+	if err != nil {
+		return err
+	}
+	defer func() {
+		_ = tx.Rollback()
+	}()
+
+	table, uniqueConstraints, err := migration.sqlschema.GetTable(ctx, sqlschema.TableName("dashboard"))
+	if err != nil {
+		return err
+	}
+
+	nameColumn := &sqlschema.Column{
+		Name:     sqlschema.ColumnName("name"),
+		DataType: sqlschema.DataTypeText,
+		Nullable: false,
+	}
+
+	// Only v2 dashboards populate this column. Existing v1 rows are left with
+	// the zero value (empty string) so v1 create/update paths can keep
+	// inserting without a name.
+	//
+	// TODO: once v1 dashboards are migrated to v2 and every row has a real
+	// name, a follow-up migration should add a unique index on
+	// (org_id, name) to enforce per-org name uniqueness.
+	sqls := migration.sqlschema.Operator().AddColumn(table, uniqueConstraints, nameColumn, nil)
+
+	for _, sql := range sqls {
+		if _, err := tx.ExecContext(ctx, string(sql)); err != nil {
+			return err
+		}
+	}
+
+	if err := tx.Commit(); err != nil {
+		return err
+	}
+
+	if err := migration.sqlschema.ToggleFKEnforcement(ctx, db, true); err != nil {
+		return err
+	}
+
+	return nil
+}
+
+func (migration *addDashboardName) Down(context.Context, *bun.DB) error {
+	return nil
+}

pkg/types/authtypes/claims.go

@@ -29,7 +29,7 @@ func NewContextWithClaims(ctx context.Context, claims Claims) context.Context {
 func ClaimsFromContext(ctx context.Context) (Claims, error) {
 	claims, ok := ctx.Value(claimsKey{}).(Claims)
 	if !ok {
-		return Claims{}, errors.New(errors.TypeUnauthenticated, errors.CodeUnauthenticated, "unauthenticated")
+		return Claims{}, errors.NewUnauthenticatedf(errors.CodeUnauthenticated, "unauthenticated")
 	}
 
 	return claims, nil
@@ -42,7 +42,7 @@ func NewContextWithAccessToken(ctx context.Context, accessToken string) context.
 func AccessTokenFromContext(ctx context.Context) (string, error) {
 	accessToken, ok := ctx.Value(accessTokenKey{}).(string)
 	if !ok {
-		return "", errors.New(errors.TypeUnauthenticated, errors.CodeUnauthenticated, "unauthenticated")
+		return "", errors.NewUnauthenticatedf(errors.CodeUnauthenticated, "unauthenticated")
 	}
 
 	return accessToken, nil
@@ -55,7 +55,7 @@ func NewContextWithAPIKey(ctx context.Context, apiKey string) context.Context {
 func APIKeyFromContext(ctx context.Context) (string, error) {
 	apiKey, ok := ctx.Value(apiKeyKey{}).(string)
 	if !ok {
-		return "", errors.New(errors.TypeUnauthenticated, errors.CodeUnauthenticated, "unauthenticated")
+		return "", errors.NewUnauthenticatedf(errors.CodeUnauthenticated, "unauthenticated")
 	}
 
 	return apiKey, nil
@@ -77,7 +77,7 @@ func (c *Claims) IsSelfAccess(id string) error {
 		return nil
 	}
 
-	return errors.New(errors.TypeForbidden, errors.CodeForbidden, "only the user/admin can access their own resource")
+	return errors.NewForbiddenf(errors.CodeForbidden, "only the user/admin can access their own resource")
 }
 
 func (c *Claims) IdentityID() string {

pkg/types/dashboardtypes/dashboard.go

@@ -33,6 +33,7 @@ type StorableDashboard struct {
 	Locked bool                  `bun:"locked,notnull,default:false"`
 	OrgID  valuer.UUID           `bun:"org_id,notnull"`
 	Source Source                `bun:"source,type:text,notnull"`
+	Name   string                `bun:"name,type:text,notnull"`
 }
 
 type Dashboard struct {