HOW WE MEASURE

Source A — Official status pages. We poll Atlassian Statuspage APIs (and RSS feeds where available) every 60 seconds. This covers ~70% of tracked services.

Source B — Synthetic probes. Real requests from at least 3 regions (us-east, eu-west, ap-southeast). We record HTTP status, latency, error class, and time-to-first-byte.

Source C — Crowdsourced reports.User-submitted "having issues?" reports with reason chips. Rate-limited by IP. Spike detection triggers badges when z-score > 3.

weariness = (downtime_minutes_7d × 1.0)
          + (wobble_minutes_7d × 0.4)
          + (user_report_spikes_7d × 2.0)
          + (major_incident_count_7d × 15)
          - (clean_streak_days × 0.5)

Higher score = wearier = ranked worse. Re-normalized daily for display as a 0–100 Weariness Index.

Wooden Spoon of the Week — worst 7-day uptime. Awarded every Monday 00:00 UTC.

Most Dramatic Outage — longest continuous single-incident downtime in the last 7/30 days.