1 / 5
A metric crosses its limit and the monitoring system sends a notification. The system ___ an alert.
-
-
-
-
Fire an alert is the standard collocation for an alert being triggered.
- fire / trigger an alert — the alert goes off
- You can also say an alert fires (intransitive)
"Shoots" and "rings off" don't fit. Example:
"When error rate exceeds 1%, the alert fires."
2 / 5
CPU usage rises above the configured 80% limit. Engineers say it ___ the threshold.
-
-
-
-
Breach a threshold means crossing the limit that triggers an alert.
- breach / cross / exceed a threshold
- The threshold is the boundary value you alert on
"Beats" and "overtakes" are not standard here. Example:
"Latency breached the 200ms threshold for five minutes."
3 / 5
It is 3am and a critical alert needs human attention now. The system ___ the on-call engineer.
-
-
-
-
Page on-call means urgently notifying the on-call engineer (from "pager").
- page the on-call / page someone
- Noun: a page, the alert that wakes you up
"Buzzes off" and "dings" are informal/wrong. Example:
"That alert pages the on-call engineer immediately."
4 / 5
An engineer sees the alert and signals they are handling it, then later marks it as fixed. They ___ it then resolve it.
-
-
-
-
Ack (acknowledge) means signalling you've seen an alert and are working on it;
resolve means closing it once fixed.
- ack an alert — acknowledge it (stops escalation)
- resolve an alert — mark it cleared
"Tick" and "flag up" aren't the on-call terms. Example:
"Ack the page within five minutes, then resolve once mitigated."
5 / 5
A monitoring tool like Prometheus periodically collects metrics by pulling them from each service. It ___ metrics.
-
-
-
-
Scrape metrics is the standard term for a monitoring system pulling metrics from endpoints.
- scrape metrics / a target — pull-based collection
- Common with Prometheus, which scrapes
/metrics endpoints on an interval
"Sucks" and "reaps" aren't used. Example:
"Prometheus scrapes each pod's metrics every 15 seconds."