Give sustained conditions time
A short burst of work does not have to become a CPU incident. Enable a threshold and elapsed duration, then use a lower clear threshold and recovery duration. CPU and memory rules are independent and disabled by default.
Fresh samples matter. Missing values or long gaps reset a pending window; an unknown reading cannot prove recovery. Learn how sustained thresholds work.
Share a sensible default
Organization policies flow to sites and machines unless a more specific override exists. Preview affected devices before a change. Rule-source labels, reset-to-inherited controls and policy history make exceptions visible.
Pause notifications during work
Maintenance starts immediately for one device and lasts for a bounded duration. Observations, actual status and history continue. Unresolved eligible notifications resume after maintenance; a recovered issue does not need a delayed warning.
Future schedules, recurring windows and bulk maintenance are not included. Read maintenance behavior.
Inspect what happened to a message
The product records notification submission status and attempt history. Only private development-inbox submission has been exercised. Production customer email and delivery/bounce tracking are pending; accepted submission is not a mailbox-delivery guarantee.