From 542c8f4525a5441df734d8ef6c80a5b83b72ebed Mon Sep 17 00:00:00 2001
From: dongmen <414110582@qq.com>
Date: Mon, 19 Jun 2023 18:47:28 +0800
Subject: [PATCH] add changefeed status alert rules

---
 metrics/alertmanager/ticdc.rules.yml | 65 +++++++++++++++++-----------
 1 file changed, 40 insertions(+), 25 deletions(-)

diff --git a/metrics/alertmanager/ticdc.rules.yml b/metrics/alertmanager/ticdc.rules.yml
index 1c5788882c1..db73b655938 100644
--- a/metrics/alertmanager/ticdc.rules.yml
+++ b/metrics/alertmanager/ticdc.rules.yml
@@ -1,6 +1,7 @@
 groups:
 - name: alert.rules
   rules:
+  # server related alter rules
   - alert: cdc_multiple_owners
     expr: sum(rate(ticdc_owner_ownership_counter[30s])) >= 2
     for: 1m
@@ -25,6 +26,19 @@ groups:
       value: '{{ $value }}'
       summary: cdc cluster has no owner for more than 10 minutes
 
+  # changefeed related alter rules
+  - alert: ticdc_changefeed_failed
+    expr: (max_over_time(ticdc_owner_status[1m]) == 2) > 0
+    for: 1m
+    labels:
+      env: ENV_LABELS_ENV
+      level: critical
+      expr: (max_over_time(ticdc_owner_status[1m]) == 2) > 0
+    annotations:
+      description: 'cluster: ENV_LABELS_ENV, instance: {{ $labels.instance }}, values: {{ $value }}'
+      value: '{{ $value }}'
+      summary: cdc changefeed failed, it can not be automatically resumed
+
   - alert: cdc_checkpoint_high_delay
     expr: ticdc_owner_checkpoint_ts_lag > 600
     for: 1m
@@ -37,62 +51,63 @@ groups:
       value: '{{ $value }}'
       summary: cdc owner checkpoint delay more than 10 minutes
 
-  - alert: tikv_cdc_min_resolved_ts_no_change_for_1m
-    expr: changes(tikv_cdc_min_resolved_ts[1m]) < 1 and ON (instance) tikv_cdc_region_resolve_status{status="resolved"} > 0
+  - alert: ticdc_sink_execution_error
+    expr: changes(ticdc_sink_execution_error[1m]) > 0
     for: 1m
     labels:
       env: ENV_LABELS_ENV
       level: warning
-      expr: changes(tikv_cdc_min_resolved_ts[1m]) < 1 and ON (instance) tikv_cdc_region_resolve_status{status="resolved"} > 0
+      expr: changes(ticdc_sink_execution_error[1m]) > 0
     annotations:
       description: 'cluster: ENV_LABELS_ENV, instance: {{ $labels.instance }}, values: {{ $value }}'
-      value: '{{ $labels.instance }}'
-      summary: tikv cdc min resolved ts no change for 1m
+      value: '{{ $value }}'
+      summary: cdc sink execution meets errors
 
-  - alert: tikv_cdc_scan_duration_seconds_more_than_10min
-    expr: histogram_quantile(0.9, rate(tikv_cdc_scan_duration_seconds_bucket{}[1m])) > 600
+  - alert: ticdc_processor_exit_with_error_count
+    expr: changes(ticdc_processor_exit_with_error_count[1m]) > 0
     for: 1m
     labels:
       env: ENV_LABELS_ENV
       level: warning
-      expr: histogram_quantile(0.9, rate(tikv_cdc_scan_duration_seconds_bucket{}[1m])) > 600
+      expr: changes(ticdc_processor_exit_with_error_count[1m]) > 0
     annotations:
       description: 'cluster: ENV_LABELS_ENV, instance: {{ $labels.instance }}, values: {{ $value }}'
       value: '{{ $value }}'
-      summary: tikv cdc scan duration seconds more than 10 min
+      summary: cdc processor exits with error
 
-  - alert: ticdc_sink_mysql_execution_error
-    expr: changes(ticdc_sink_mysql_execution_error[1m]) > 0
+  - alert: ticdc_changefeed_meet_error
+    expr: (max_over_time(ticdc_owner_status[1m]) == 1) > 0
     for: 1m
     labels:
       env: ENV_LABELS_ENV
       level: warning
-      expr: changes(ticdc_sink_mysql_execution_error[1m]) > 0
+      expr: (max_over_time(ticdc_owner_status[1m]) == 1) > 0
     annotations:
       description: 'cluster: ENV_LABELS_ENV, instance: {{ $labels.instance }}, values: {{ $value }}'
       value: '{{ $value }}'
-      summary: cdc sink mysql execution meets errors
-
-  - alert: ticdc_processor_exit_with_error_count
-    expr: changes(ticdc_processor_exit_with_error_count[1m]) > 0
+      summary: cdc changefeed meet error
+  
+  # tikv related alter rules
+  - alert: tikv_cdc_min_resolved_ts_no_change_for_1m
+    expr: changes(tikv_cdc_min_resolved_ts[1m]) < 1 and ON (instance) tikv_cdc_region_resolve_status{status="resolved"} > 0
     for: 1m
     labels:
       env: ENV_LABELS_ENV
-      level: critical
-      expr: changes(ticdc_processor_exit_with_error_count[1m]) > 0
+      level: warning
+      expr: changes(tikv_cdc_min_resolved_ts[1m]) < 1 and ON (instance) tikv_cdc_region_resolve_status{status="resolved"} > 0
     annotations:
       description: 'cluster: ENV_LABELS_ENV, instance: {{ $labels.instance }}, values: {{ $value }}'
-      value: '{{ $value }}'
-      summary: cdc processor exits with error
+      value: '{{ $labels.instance }}'
+      summary: tikv cdc min resolved ts no change for 1m
 
-  - alert: ticdc_memory_abnormal
-    expr: go_memstats_heap_alloc_bytes{job="ticdc"} > 1e+10
+  - alert: tikv_cdc_scan_duration_seconds_more_than_10min
+    expr: histogram_quantile(0.9, rate(tikv_cdc_scan_duration_seconds_bucket{}[1m])) > 600
     for: 1m
     labels:
       env: ENV_LABELS_ENV
       level: warning
-      expr: go_memstats_heap_alloc_bytes{job="ticdc"} > 1e+10
+      expr: histogram_quantile(0.9, rate(tikv_cdc_scan_duration_seconds_bucket{}[1m])) > 600
     annotations:
-      description: 'cluster: ENV_LABELS_ENV, instance: {{ $labels.instance }}, values:{{ $value }}'
+      description: 'cluster: ENV_LABELS_ENV, instance: {{ $labels.instance }}, values: {{ $value }}'
       value: '{{ $value }}'
-      summary: TiCDC heap memory usage is over 10 GB
+      summary: tikv cdc scan duration seconds more than 10 min
\ No newline at end of file