Monitor visit marker

CHANGELOG.md

@@ -2,6 +2,7 @@
 
 ## master
 * [ENHANCEMENT] Add bigger tenants and configure default compactor tenant shards
+* [ENHANCEMENT] Add alert `CortexCompactorWriteVisitMarkerIsFailing` to monitor compactors
 
 ## 1.17.1 / 2024-10-23
 * [CHANGE] Use cortex v1.17.1

cortex-mixin/alerts/compactor.libsonnet

@@ -102,6 +102,22 @@
             ||| % $._config,
           },
         },
+        {
+          // Alert if compactor are not able to update the visit-marker.
+          alert: 'CortexCompactorBlockVisitMarkerIsFailing',
+          'for': '2h',
+          expr: |||
+            sum(increase(cortex_compactor_block_visit_marker_write_failed{job=~".+/%(compactor)s"}[2h]))>0
+          ||| % $._config.job_names,
+          labels: {
+            severity: 'critical'
+          },
+          annotations: {
+            message: |||
+              Cortex compactors are not able to update the visit marker, double check logs to see what is happening
+            |||
+          }
+        }
       ],
     },
   ],

cortex-mixin/docs/playbooks.md

@@ -379,6 +379,17 @@ How to **investigate**:
 - Ensure ingesters are successfully shipping blocks to the storage
 - Look for any error in the compactor logs
 
+### CortexCompactorWriteVisitMarkerIsFailing
+
+Only applies to compactors when using shuffle sharding.
+This alert fires if the compactor is not able to update the visit marker across all tenants.
+The marker file is a very small json file that should never have any problems getting updated.
+
+How to **investigate**:
+- Verify the logs for the compactors, they should show the exact reason
+- If you see the `context canceled` or any other timeouts in the logs,
+consider increasing `-compactor.compaction-visit-marker-timeout` and `-compactor.compaction-visit-marker-file-update-interval`.
+
 ### CortexCompactorHasNotSuccessfullyRunCompaction
 
 This alert fires if the compactor is not able to successfully compact all discovered compactable blocks (across all tenants).