Alert (Monitoring)

An alert is a notification generated when a monitoring system decides that a condition needs human attention. The condition might be a failed HTTP check, an error rate above a threshold, or an expiring certificate. Alerts are delivered by email, SMS, push, chat message, phone call or a paging service.

Why it matters

Monitoring that nobody sees is only a record. Alerts close the loop between detection and response, so the time between a failure and a person acting on it is short. Quality matters more than quantity: every alert should be urgent, actionable and clear about what is wrong. Alerts that fail those tests train people to ignore them.

Example

[CRITICAL] shop.example.com
HTTP check failed from 3 of 3 locations
Status: 503 Service Unavailable, started 09:12 UTC
Runbook: https://wiki.example.com/runbooks/shop-503

This includes what, where, since when and what to do next.

Common mistakes

• Alerting on every anomaly instead of conditions that need action.
• Sending everything to one shared inbox no one watches.
• Omitting context such as affected service, severity and a link to a runbook.
• Firing on a single failed check instead of confirming from several locations.

Related

• Alert fatigue
• Threshold alert
• Escalation policy
• False positive

यह पेज सामान्य जानकारी है, जिसे हमारी संपादकीय टीम ने लिखा है और दस्तावेज़ीकरण बदलने पर अपडेट किया जाता है। सर्वर और प्रदाता अलग-अलग होते हैं, इसलिए बदलाव पहले सुरक्षित तरीके से जाँचें। देखें: संपादकीय नीति और अस्वीकरण.