Q: We deploy nightly and it pages everyone. How do I set up maintenance windows right?
A: Suppressing alerts during planned work is good hygiene, but do it carefully so you don't blind yourself.
— Create recurring maintenance windows that match your real deploy schedule, with a 5-minute buffer on each end.
— Suppress alerting, but keep recording data. You still want the uptime history accurate.
— Decide whether the window counts against your SLA. Planned maintenance usually shouldn't, but write it down.
— Auto-post a 'scheduled maintenance' banner to your status page during the window.
— Set a hard max duration so a forgotten window doesn't hide a real outage for hours.
Follow-up: if a deploy runs long and the window closes mid-outage, you want the alert to fire then, not stay muted. Cap the window, don't make it open-ended.
Got a question? Drop it in the comments.
Pingback Clinic
@PingbackClinic
Q: We deploy nightly and it pages everyone. How do I set up maintenance windows right?
Этот пост опубликован в Telegram-канале Pingback Clinic. Подписаться можно по ссылке: @PingbackClinic.