Why Assistant Watchers
A watcher is an always-on agent. You describe what healthy looks like for a service or failure mode. Assistant turns that intent into checks, then the watcher scans telemetry on a schedule and raises concern when a run finds something notable.
A watcher can stay quiet when signals look normal. When it does raise concern, it can notify Slack, Microsoft Teams, or a webhook endpoint, send findings to Grafana Alerting, and launch an Assistant Investigation for critical findings.
Use a watcher for these kinds of work:
- Recurring service health checks
- Noisy-but-important failure modes
- Follow-up monitoring after an incident
A watcher can also help where an alert rule is hard to write. For example, a new service can benefit from a flexible watcher while its normal ranges are still changing.
In this path you will:
- Open Watchers and create a watcher with a narrow scope
- Calibrate checks against your telemetry
- Trigger Run now and pause the automatic runs with the Active switch
- Check the watcher’s verdict and review a run report
In the next milestone, you’ll open Watchers and create one.