Cleanup code to have only query activity

Author: sangeetashivaji

clickhouse/assets/configuration/spec.yaml

@@ -76,7 +76,7 @@ files:
         example: True
     - name: dbm
       description: |
-        Enable Database Monitoring (DBM) to collect query samples and execution plans.
+        Enable Database Monitoring (DBM) to collect query activity, metrics, and samples.
         This feature provides deep observability into query performance.
       value:
         type: boolean
@@ -95,63 +95,37 @@ files:
       value:
         type: boolean
         example: false
-    - name: query_samples
-      description: Configure collection of query samples
+    - name: query_activity
+      description: Configure collection of database activity snapshots from system.processes
       options:
         - name: enabled
           description: |
-            Enable collection of query samples. Requires `dbm: true`.
+            Enable collection of database activity snapshots. Requires `dbm: true`.
+            Activity snapshots show currently executing queries and connection counts.
           value:
             type: boolean
             example: true
         - name: collection_interval
           description: |
-            Set the query sample collection interval (in seconds). Each collection involves a single query to
+            Set the activity snapshot collection interval (in seconds). Each collection involves a single query to
             `system.processes` to capture currently executing queries.
           value:
             type: number
             example: 1
-        - name: samples_per_hour_per_query
-          description: |
-            Set the rate limit for how many query sample events will be ingested per hour per normalized query.
-          value:
-            type: integer
-            example: 15
-        - name: seen_samples_cache_maxsize
+        - name: payload_row_limit
           description: |
-            Set the max size of the cache used for the samples_per_hour_per_query rate limit. This should be increased
-            for databases with a very large number of unique normalized queries which exceed the cache's limit.
+            Set the maximum number of active sessions to include in each activity snapshot.
           value:
             type: integer
-            example: 10000
+            example: 1000
         - name: run_sync
           hidden: true
           description: |
-            Run the query samples collection synchronously. This is useful for testing purposes, but should not be used
+            Run the activity collection synchronously. This is useful for testing purposes, but should not be used
             in production as it can block the main thread and cause performance issues.
           value:
             type: boolean
             example: false
-        - name: activity_enabled
-          description: |
-            Enable collection of database activity snapshots. Requires `dbm: true`.
-          value:
-            type: boolean
-            example: true
-        - name: activity_collection_interval
-          hidden: true
-          description: |
-            Set the activity snapshot collection interval (in seconds). This number cannot be smaller than
-            the `query_samples` configured collection_interval.
-          value:
-            type: number
-            example: 10
-        - name: activity_max_rows
-          description: |
-            Set the maximum number of active sessions to include in each activity snapshot.
-          value:
-            type: integer
-            example: 1000
     - name: query_metrics
       description: Configure collection of query metrics
       options:

clickhouse/datadog_checks/clickhouse/clickhouse.py

@@ -15,7 +15,7 @@
 from .__about__ import __version__
 from .completed_query_samples import ClickhouseCompletedQuerySamples
 from .config import build_config
-from .statement_samples import ClickhouseStatementSamples
+from .statement_activity import ClickhouseStatementActivity
 from .statements import ClickhouseStatementMetrics
 from .utils import ErrorSanitizer
 
@@ -91,11 +91,11 @@ def _init_dbm_components(self):
         else:
             self.statement_metrics = None
 
-        # Initialize query samples (from system.processes - analogous to pg_stat_activity)
-        if self._config.dbm and self._config.query_samples.enabled:
-            self.statement_samples = ClickhouseStatementSamples(self, self._config.query_samples)
+        # Initialize query activity (from system.processes - analogous to pg_stat_activity)
+        if self._config.dbm and self._config.query_activity.enabled:
+            self.statement_activity = ClickhouseStatementActivity(self, self._config.query_activity)
         else:
-            self.statement_samples = None
+            self.statement_activity = None
 
         # Initialize completed query samples (from system.query_log - completed queries)
         if self._config.dbm and self._config.completed_query_samples.enabled:
@@ -182,9 +182,9 @@ def check(self, _):
         if self.statement_metrics:
             self.statement_metrics.run_job_loop(self.tags)
 
-        # Run statement samples if DBM is enabled (from system.processes)
-        if self.statement_samples:
-            self.statement_samples.run_job_loop(self.tags)
+        # Run query activity collection if DBM is enabled (from system.processes)
+        if self.statement_activity:
+            self.statement_activity.run_job_loop(self.tags)
 
         # Run completed query samples if DBM is enabled (from system.query_log)
         if self.completed_query_samples: