System Pulse
System Pulse introduces a new /pulse page that shows what needs attention across your entire stack.
Instead of checking multiple pages, Pulse brings together issues from Sonarr, Radarr, Seerr, cache systems, and background jobs into a single prioritized feed.
All signals in one place — critical, warning, and info items across your system
What It Shows
Pulse surfaces:
- Critical issues — unreachable instances, service failures
- Warnings — degraded health, stale data, recent failures
- Info signals — opportunities like quality upgrades
Each item clearly answers:
- what is happening
- why it matters now
- where to go next
Expand sections to see all items — action links navigate directly to the relevant page
Why It Matters
Before Pulse, you had to manually check:
- Dashboard → Statistics → Library → Settings → Hunting → Queue Cleaner → TRaSH Guides
Now: one page, one view, one prioritized list.
Signals Included
| Signal | Severity | Source |
|---|---|---|
| ARR health issues | critical/warning | Sonarr, Radarr, Prowlarr, Lidarr, Readarr |
| Disk space >80%/>90% | warning/critical | ARR instances (deduplicated) |
| Instance unreachable | critical | Failed SDK health calls |
| Seerr circuit breaker | critical/warning | OPEN/HALF_OPEN state |
| Cache staleness | warning | Plex/Tautulli (error or >12h stale) |
| Validation health | critical/warning | Integration health registry |
| Hunt failures | warning | Last 24h errors |
| Queue cleaner failures | warning | Last 24h errors |
| TRaSH sync failures | warning | Most recent per instance |
| Quality cutoff | info | Library items below cutoff |
Clean, Trustworthy Signals
- No false positives — signals only appear when services are configured and issues exist
- No cross-user data leakage — all queries scoped by userId
- Full incognito mode support — instance names and health messages anonymized
- Accurate action links — every item navigates to the correct page
Notes
- No schema changes — no database migration required
- Signals computed on request with 60-second per-user cache
- Frontend polls every 120 seconds
Full changelog: CHANGELOG.md