Alerts about each are held back for this long; anything still wrong when it ends alerts then. One already under maintenance has its time restarted.
setReason(e.target.value)} placeholder="e.g. Firmware update" /> {canSilenceDependents && ( )}Says that somebody has seen them. The alerts stay, and are still counted by /metrics, but they are no longer ones nobody has looked at. One that was already acknowledged has its note replaced.
When the server they run on is down, their own offline alerts are folded under its alert instead of adding to the noise. A change that would make a loop is left out.
Whatever is left as it is stays the way each one has it.