Alerts

Monitoring only helps if it tells you in time. Lookout’s alerting is designed to warn you the moment a server crosses a threshold or goes silent — and to stay quiet otherwise.

Status: the dashboard computes health today (ok / warning / critical / stale). Notification delivery (email/Slack/webhook), dedup, and escalation are the next phase — this page describes how they work so you can plan for them.

What triggers an alert

• A server crosses a threshold (e.g. disk ≥ 90%, memory ≥ 90%).
• A critical service stops running.
• A server goes stale — no report for 5 minutes — which usually means it’s down.
• Any custom plugin returns a warning or critical result.

Channels

• Email — a plain-English summary of what changed and what to do.
• Slack — a message to the channel of your choice.
• Webhook — POST the event anywhere (PagerDuty, Opsgenie, your own handler).

Designed to avoid noise

• Dedup — one alert per problem, not one per check run.
• Recovery — you’re told when it clears, too.
• Escalation — if a critical isn’t acknowledged, escalate to the next contact.

Gate your pipeline

The agent can also fail a CI/CD job if a previously-passing check regressed — so you catch drift before you ship:

lookout-agent run --once --fail-on-drift   # exits non-zero on a regression

(--fail-on-drift is part of the monitoring roadmap.)