你又一次被叫醒,卻什麼事都沒有。 Polylane 讀每一則警示,並讓雜訊安靜下來。
加入候補名單 Issues Elevated error rate on payments-api Timeline
Elevated error rate on payments-api
High
Incident
·Detected from a Datadog alert
OverviewMetricsLogsTracesTimelinePropertiesInvestigation
- 02:14:07 Alert received
Datadog · payments-api error rate · occurrence #1
- 02:14:09 Triage started
Fetching the monitor query and metrics 30 min before, 5 min after the firing.
- 02:14:58 Triaged as an incident
Severity high. Error rate 41× the 30-minute baseline, clustered on POST /v1/charges; deploy f3a9c12 landed 18 minutes before the firing.
- 02:15:02 Investigation started
3 hypotheses · 3 passes each
- 02:31:44 Alert fired again (occurrence #2)
Folded into this issue after 16 quiet minutes; the running investigation was nudged.
- 03:06:12 Alert recovered
Provider recovery received. Issue resolved.
雜訊到此為止。 每一則警示都經過分流。
-
風暴收攏
同一則呼叫重複一百次,併成一則。
-
認得出重複
再次觸發會落在它所屬的問題上,而不是開一個新的,即使你的供應商重新分組了它。
-
依資料判斷
代理取出你監控背後真正的查詢,以及觸發前後的指標,再與正常狀態比較。
-
事件或雜訊
雜訊被歸檔。事件會自動被調查,而當你的供應商恢復時,問題也會跟著關閉。