Changed the alert rule of TiKV_thread_apply_worker_cpu_seconds (#7068) (

#7347)
pingcap · Dec 29, 2021 · d99804e · d99804e
1 parent 7054cab
commit d99804e
Showing 1 changed file with 2 additions and 2 deletions.
diff --git a/alert-rules.md b/alert-rules.md
@@ -595,11 +595,11 @@ This section gives the alert rules for the TiKV component.
 
 * Alert rule:
 
-    `sum(rate(tikv_thread_cpu_seconds_total{name="apply_worker"}[1m])) by (instance) > 1.8`
+    `max(rate(tikv_thread_cpu_seconds_total{name=~"apply_.*"}[1m])) by (instance) > 0.9`
 
 * Description:
 
-    The pressure on the apply Raft log thread is too high. It is often caused by a burst of writes.
+    The apply Raft log thread is under great pressure and is approaching or has exceeded its limit. This is often caused by a burst of writes.
 
 ### Warning-level alerts