SigSentrySigSentry

Watchdog

Proactive monitoring — set rules on your log sources, get notified when conditions are met, optionally auto-run a full diagnosis

Watchdog turns SigSentry from a reactive tool ("something's wrong, run an analysis") into a proactive one ("alert me when something looks off, and skip the manual step"). You define rules against your log sources; Watchdog watches around the clock and notifies you — or runs a full analysis automatically — when a rule fires.

When to use it

  • You want to be paged before customers report an issue
  • You have known error patterns you'd like to catch automatically
  • You're triaging incidents from log dashboards manually and want to skip that step

How it works

Quick walkthrough

Pick a log source

Open Project → Watchdog → New rule. Pick the log source you want to monitor. Watchdog operates per-source, so a project with multiple sources gets multiple rules.

Choose a rule type

Error count, error rate, pattern match, or spike detection. Each fits a different operational shape — see Rule types.

Set thresholds and lookback

Watchdog evaluates each rule against the last N minutes of logs. The default lookback is 30 minutes — long enough to smooth out brief blips, short enough to catch real problems quickly.

Pick alert behavior

Notify-only sends a message to your channels. Auto-analyze runs a full diagnosis when the rule fires and posts the result. See Alert behavior.

Dry-run, then enable

Always dry-run before enabling. It replays the rule against recent logs and shows you exactly how often it would have fired — saves you from a noisy rule going live.

Plan limits

PlanWatchdog rules per project
Starter0 (Watchdog disabled)
Pro3
Business10
EnterpriseUnlimited

Each enabled rule consumes scheduled checks against your log source. Auto-analyze rules also consume your monthly analysis quota — use the cost calculator to estimate.