test: histogram percentile integration tests

.devenv/docker/clickhouse/compose.yaml

@@ -1,4 +1,22 @@
 services:
+  init-clickhouse:
+    image: clickhouse/clickhouse-server:25.5.6
+    container_name: init-clickhouse
+    command:
+      - bash
+      - -c
+      - |
+        version="v0.0.1"
+        node_os=$$(uname -s | tr '[:upper:]' '[:lower:]')
+        node_arch=$$(uname -m | sed s/aarch64/arm64/ | sed s/x86_64/amd64/)
+        echo "Fetching histogram-binary for $${node_os}/$${node_arch}"
+        cd /tmp
+        wget -O histogram-quantile.tar.gz "https://github.com/SigNoz/signoz/releases/download/histogram-quantile%2F$${version}/histogram-quantile_$${node_os}_$${node_arch}.tar.gz"
+        tar -xvzf histogram-quantile.tar.gz
+        mv histogram-quantile /var/lib/clickhouse/user_scripts/histogramQuantile
+    restart: on-failure
+    volumes:
+      - ${PWD}/fs/tmp/var/lib/clickhouse/user_scripts/:/var/lib/clickhouse/user_scripts/
   clickhouse:
     image: clickhouse/clickhouse-server:25.5.6
     container_name: clickhouse
@@ -7,6 +25,7 @@ services:
       - ${PWD}/fs/etc/clickhouse-server/users.d/users.xml:/etc/clickhouse-server/users.d/users.xml
       - ${PWD}/fs/tmp/var/lib/clickhouse/:/var/lib/clickhouse/
       - ${PWD}/fs/tmp/var/lib/clickhouse/user_scripts/:/var/lib/clickhouse/user_scripts/
+      - ${PWD}/../../../deploy/common/clickhouse/custom-function.xml:/etc/clickhouse-server/custom-function.xml
     ports:
       - '127.0.0.1:8123:8123'
       - '127.0.0.1:9000:9000'
@@ -22,7 +41,10 @@ services:
       timeout: 5s
       retries: 3
     depends_on:
-      - zookeeper
+      init-clickhouse:
+        condition: service_completed_successfully
+      zookeeper:
+        condition: service_healthy
     environment:
       - CLICKHOUSE_SKIP_USER_SETUP=1
   zookeeper:

.devenv/docker/clickhouse/fs/etc/clickhouse-server/config.d/config.xml

@@ -44,4 +44,5 @@
         <shard>01</shard>
         <replica>01</replica>
     </macros>
+    <user_defined_executable_functions_config>*function.xml</user_defined_executable_functions_config>
 </clickhouse>

tests/integration/fixtures/clickhouse.py

@@ -75,6 +75,8 @@ def clickhouse(
                 </cluster>
             </remote_servers>
 
+            <user_defined_executable_functions_config>*function.xml</user_defined_executable_functions_config>
+
             <distributed_ddl>
                 <path>/clickhouse/task_queue/ddl</path>
                 <profile>default</profile>
@@ -117,17 +119,74 @@ def clickhouse(
         </clickhouse>
         """
 
+        custom_function_config = """
+        <functions>
+            <function>
+                <type>executable</type>
+                <name>histogramQuantile</name>
+                <return_type>Float64</return_type>
+                <argument>
+                    <type>Array(Float64)</type>
+                    <name>buckets</name>
+                </argument>
+                <argument>
+                    <type>Array(Float64)</type>
+                    <name>counts</name>
+                </argument>
+                <argument>
+                    <type>Float64</type>
+                    <name>quantile</name>
+                </argument>
+                <format>CSV</format>
+                <command>./histogramQuantile</command>
+            </function>
+        </functions>
+        """
+
         tmp_dir = tmpfs("clickhouse")
         cluster_config_file_path = os.path.join(tmp_dir, "cluster.xml")
         with open(cluster_config_file_path, "w", encoding="utf-8") as f:
             f.write(cluster_config)
 
+        custom_function_file_path = os.path.join(tmp_dir, "custom-function.xml")
+        with open(custom_function_file_path, "w", encoding="utf-8") as f:
+            f.write(custom_function_config)
+
         container.with_volume_mapping(
             cluster_config_file_path, "/etc/clickhouse-server/config.d/cluster.xml"
         )
+        container.with_volume_mapping(
+            custom_function_file_path,
+            "/etc/clickhouse-server/custom-function.xml",
+        )
         container.with_network(network)
         container.start()
 
+        # Download and install the histogramQuantile binary
+        wrapped = container.get_wrapped_container()
+        exit_code, output = wrapped.exec_run(
+            [
+                "bash",
+                "-c",
+                (
+                    'version="v0.0.1" && '
+                    'node_os=$(uname -s | tr "[:upper:]" "[:lower:]") && '
+                    'node_arch=$(uname -m | sed s/aarch64/arm64/ | sed s/x86_64/amd64/) && '
+                    "cd /tmp && "
+                    'wget -O histogram-quantile.tar.gz "https://github.com/SigNoz/signoz/releases/download/histogram-quantile%2F${version}/histogram-quantile_${node_os}_${node_arch}.tar.gz" && '
+                    "tar -xzf histogram-quantile.tar.gz && "
+                    "mkdir -p /var/lib/clickhouse/user_scripts && "
+                    "mv histogram-quantile /var/lib/clickhouse/user_scripts/histogramQuantile && "
+                    "chmod +x /var/lib/clickhouse/user_scripts/histogramQuantile"
+                ),
+            ],
+        )
+        if exit_code != 0:
+            logger.warning(
+                "Failed to install histogramQuantile binary: %s",
+                output.decode(),
+            )
+
         connection = clickhouse_connect.get_client(
             user=container.username,
             password=container.password,

tests/integration/src/querier/08_metrics_histogram.py

@@ -372,3 +372,153 @@ def test_histogram_count_no_param(
             values[1]["value"] == first_values[le]
         )  ## to keep parallel to the cumulative test cases, first_value refers to the value at 10:02
         assert values[-1]["value"] == last_values[le]
+
+@pytest.mark.parametrize(
+    "space_agg, zeroth_value, first_value, last_value",
+    [
+        ("p50", 500, 818.182, 550.725),
+        ("p75", 750, 3000, 826.087),
+        ("p90", 900, 6400, 991.304),
+        ("p95", 950, 8000, 4200),
+        ("p99", 990, 8000, 8000),
+    ],
+)
+def test_histogram_percentile_for_all_services(
+    signoz: types.SigNoz,
+    create_user_admin: None,  # pylint: disable=unused-argument
+    get_token: Callable[[str, str], str],
+    insert_metrics: Callable[[List[Metrics]], None],
+    space_agg: str,
+    zeroth_value: float,
+    first_value: float,
+    last_value: float,
+) -> None:
+    now = datetime.now(tz=timezone.utc).replace(second=0, microsecond=0)
+    start_ms = int((now - timedelta(minutes=65)).timestamp() * 1000)
+    end_ms = int(now.timestamp() * 1000)
+    metric_name = f"test_{space_agg}_bucket"
+
+    metrics = Metrics.load_from_file(
+        FILE,
+        base_time=now - timedelta(minutes=60),
+        metric_name_override=metric_name,
+    )
+    insert_metrics(metrics)
+
+    token = get_token(USER_ADMIN_EMAIL, USER_ADMIN_PASSWORD)
+    query = build_builder_query(
+        "A",
+        metric_name,
+        "doesnotreallymatter",
+        space_agg,
+    )
+
+    response = make_query_request(signoz, token, start_ms, end_ms, [query])
+    assert response.status_code == HTTPStatus.OK
+
+    data = response.json()
+    result_values = sorted(get_series_values(data, "A"), key=lambda x: x["timestamp"])
+    assert len(result_values) == 60
+    assert result_values[0]["value"] == zeroth_value
+    assert result_values[1]["value"] == first_value
+    assert result_values[-1]["value"] == last_value
+
+@pytest.mark.parametrize(
+    "space_agg, first_value, last_value",
+    [
+        ("p50", 818.182, 550.725),
+        ("p75", 3000, 826.087),
+        ("p90", 6400, 991.304),
+        ("p95", 8000, 4200),
+        ("p99", 8000, 8000),
+    ],
+)
+def test_histogram_percentile_for_cumulative_service(
+    signoz: types.SigNoz,
+    create_user_admin: None,  # pylint: disable=unused-argument
+    get_token: Callable[[str, str], str],
+    insert_metrics: Callable[[List[Metrics]], None],
+    space_agg: str,
+    first_value: float,
+    last_value: float,
+) -> None:
+    now = datetime.now(tz=timezone.utc).replace(second=0, microsecond=0)
+    start_ms = int((now - timedelta(minutes=65)).timestamp() * 1000)
+    end_ms = int(now.timestamp() * 1000)
+    metric_name = f"test_{space_agg}_cumulative_bucket"
+
+    metrics = Metrics.load_from_file(
+        FILE,
+        base_time=now - timedelta(minutes=60),
+        metric_name_override=metric_name,
+    )
+    insert_metrics(metrics)
+
+    token = get_token(USER_ADMIN_EMAIL, USER_ADMIN_PASSWORD)
+    query = build_builder_query(
+        "A",
+        metric_name,
+        "doesnotreallymatter",
+        space_agg,
+        filter_expression='service = "api"',
+    )
+
+    response = make_query_request(signoz, token, start_ms, end_ms, [query])
+    assert response.status_code == HTTPStatus.OK
+
+    data = response.json()
+    result_values = sorted(get_series_values(data, "A"), key=lambda x: x["timestamp"])
+    assert len(result_values) == 59
+    assert result_values[0]["value"] == first_value
+    assert result_values[-1]["value"] == last_value
+
+@pytest.mark.parametrize(
+    "space_agg, zeroth_value, first_value, last_value",
+    [
+        ("p50", 500, 818.182, 550.725),
+        ("p75", 750, 3000, 826.087),
+        ("p90", 900, 6400, 991.304),
+        ("p95", 950, 8000, 4200),
+        ("p99", 990, 8000, 8000),
+    ],
+)
+def test_histogram_percentile_for_delta_service(
+    signoz: types.SigNoz,
+    create_user_admin: None,  # pylint: disable=unused-argument
+    get_token: Callable[[str, str], str],
+    insert_metrics: Callable[[List[Metrics]], None],
+    space_agg: str,
+    zeroth_value: float,
+    first_value: float,
+    last_value: float,
+) -> None:
+    now = datetime.now(tz=timezone.utc).replace(second=0, microsecond=0)
+    start_ms = int((now - timedelta(minutes=65)).timestamp() * 1000)
+    end_ms = int(now.timestamp() * 1000)
+    metric_name = f"test_{space_agg}_bucket"
+
+    metrics = Metrics.load_from_file(
+        FILE,
+        base_time=now - timedelta(minutes=60),
+        metric_name_override=metric_name,
+    )
+    insert_metrics(metrics)
+
+    token = get_token(USER_ADMIN_EMAIL, USER_ADMIN_PASSWORD)
+    query = build_builder_query(
+        "A",
+        metric_name,
+        "doesnotreallymatter",
+        space_agg,
+        filter_expression='service = "web"',
+    )
+
+    response = make_query_request(signoz, token, start_ms, end_ms, [query])
+    assert response.status_code == HTTPStatus.OK
+
+    data = response.json()
+    result_values = sorted(get_series_values(data, "A"), key=lambda x: x["timestamp"])
+    assert len(result_values) == 60
+    assert result_values[0]["value"] == zeroth_value
+    assert result_values[1]["value"] == first_value
+    assert result_values[-1]["value"] == last_value