metrics alert: fix apply cpu alert (tikv#11672)

Signed-off-by: zhangjinpeng1987 <[email protected]> Co-authored-by: Ti Chi Robot <[email protected]>
3pointer · Dec 27, 2021 · 1bd88a8 · 1bd88a8
1 parent 597aa14
commit 1bd88a8
Showing 1 changed file with 2 additions and 2 deletions.
diff --git a/metrics/alertmanager/tikv.rules.yml b/metrics/alertmanager/tikv.rules.yml
@@ -158,12 +158,12 @@ groups:
       summary: TiKV scheduler latch wait duration seconds more than 1s
 
   - alert: TiKV_thread_apply_worker_cpu_seconds
-    expr: sum(rate(tikv_thread_cpu_seconds_total{name="apply_worker"}[1m])) by (instance) > 1.8
+    expr: max(rate(tikv_thread_cpu_seconds_total{name="apply_.*"}[1m])) by (instance) > 0.9
     for: 1m
     labels:
       env: ENV_LABELS_ENV
       level: critical
-      expr:  sum(rate(tikv_thread_cpu_seconds_total{name="apply_worker"}[1m])) by (instance) > 1.8
+      expr:  max(rate(tikv_thread_cpu_seconds_total{name="apply_.*"}[1m])) by (instance) > 0.9
     annotations:
       description: 'cluster: ENV_LABELS_ENV, type: {{ $labels.type }}, instance: {{ $labels.instance }}, values: {{ $value }}'
       value: '{{ $value }}'