Adapter Health Lifecycle — DOWN Streaks, Alerts and Recovery

Adapter Health Lifecycle — DOWN Streaks, Alerts and Recovery A lifecycle diagram generated by Archify. 01 / Health phases 02 / Alert suppression 03 / Outcomes Healthy · 0 consecutive DOWN · Health phases · green 01 Healthy 0 consecutive DOWN green Degraded · 1–2 consecutive DOWN · Health phases · amber 02 Degraded 1–2 consecutive DOWN amber Alert active · 3rd DOWN raises one alert · Health phases · red 03 Alert active 3rd DOWN raises one alert red Streak continues · no second alert · Alert suppression · 4th+ DOWN Streak continues no second alert 4th+ DOWN UP UP The alert closes; a later streak alerts again Legend start active state waiting failure / exit

Threshold

  • • Two DOWNs stay amber without an alert
  • • The third raises exactly one alert per streak
  • • AdapterHealth.observe returns the next state plus an optional alert

Suppression

  • • A 4th or later DOWN keeps counting the streak
  • • alertActive stays true, so no second alert is derived
  • • DEGRADED observations leave both counters untouched

Recovery and replays

  • • Any UP resets the streak and closes the alert
  • • Alert ids derive from the trigger, so replays never duplicate one
  • • The aggregate is immutable: every observation returns a new state