fix(clickstack): bound ClickHouse server log and system log table growth on the data volume

.changeset/bound-clickhouse-log-growth.md

@@ -0,0 +1,9 @@
+---
+"helm-charts": minor
+---
+
+fix(clickstack): bound ClickHouse server log and system log table growth on the data volume
+
+The ClickHouse operator defaults to `logger.level: trace` with 50 x 1000M rotated log files, and enables `query_log`, `part_log`, `text_log`, `metric_log` and `asynchronous_metric_log` without a TTL. Both are written to the ClickHouse data volume, so the chart's 10Gi default fills up within days of light use, after which the OTel collector drops every batch. The chart now sets `clickhouse.cluster.spec.settings.logger` to `information` / `100M` / 10 files and adds a 7 day TTL to those five system tables via `extraConfig`. Override either block in your values to keep more history.
+
+On upgrade, ClickHouse recreates each system table whose TTL changed and keeps the old rows in `system.<table>_0`; drop those once you no longer need them.

charts/clickstack/tests/clickhouse-service_test.yaml

@@ -9,3 +9,18 @@ tests:
           of: ClickHouseCluster
       - isNotNull:
           path: spec.settings.extraConfig
+
+  - it: should bound server log and system log table growth on the data volume
+    asserts:
+      - equal:
+          path: spec.settings.logger.level
+          value: information
+      - equal:
+          path: spec.settings.logger.count
+          value: 10
+      - equal:
+          path: spec.settings.extraConfig.text_log.ttl
+          value: event_date + INTERVAL 7 DAY DELETE
+      - equal:
+          path: spec.settings.extraConfig.metric_log.ttl
+          value: event_date + INTERVAL 7 DAY DELETE

charts/clickstack/values.yaml

@@ -462,10 +462,27 @@ clickhouse:
               profile: default
               grants:
                 - query: "GRANT SELECT,INSERT,CREATE,SHOW ON default.*"
+        # The operator defaults to level trace with 50 x 1000M log files, written
+        # to the data volume. That fills the 10Gi default volume within days.
+        logger:
+          level: information
+          size: 100M
+          count: 10
         extraConfig:
           max_connections: 4096
           keep_alive_timeout: 64
           max_concurrent_queries: 100
+          # System log tables also live on the data volume and have no TTL by default.
+          query_log:
+            ttl: event_date + INTERVAL 7 DAY DELETE
+          part_log:
+            ttl: event_date + INTERVAL 7 DAY DELETE
+          text_log:
+            ttl: event_date + INTERVAL 7 DAY DELETE
+          metric_log:
+            ttl: event_date + INTERVAL 7 DAY DELETE
+          asynchronous_metric_log:
+            ttl: event_date + INTERVAL 7 DAY DELETE
 
 # Additional Kubernetes manifests to deploy alongside the chart.
 # Each entry is a complete Kubernetes object. Values are rendered through