ServiceRadar v1.4.62

Capacity forecasting now watches each disk mount point on its
own, core health checks raise an incident when they go
unhealthy, and a crossing far beyond the forecast horizon is
labeled as such instead of as a history-cap skip.

Whats New

1.4.62

• Capacity: disk usage is forecast per mount point. A new
  hourly continuous aggregate (timeseries_metrics_disk_hourly)
  keeps sysmon.disk per device, series and mount point, so a
  data volume filling toward 100 percent is no longer averaged
  away behind a flat root filesystem. Exposed as SRQL entity
  timeseries_metric_disk_hourly and JSON:API
  /timeseries_metrics_disk_hourly (PR #415).
• Core: a core health check transitioning to unhealthy now
  opens one critical incident per entity and recovers it on
  the healthy transition, via the seeded
  core_health_state_change_events and
  core_health_check_unhealthy rules. Health log lines carry a
  structured health attribute block (PR #413).
• Capacity: a capped crossing that lies beyond the 90 day
  horizon is recorded as outside_forecast_horizon instead of
  exhaustion_beyond_history_cap, with the crossing kept in
  diagnostics (PR #412).

Upgrade Notes

• New migration 20260912120000 creates the disk hourly
  aggregate WITH NO DATA and registers its refresh and
  retention policies. The chart default for
  core.migrations.expectedVersion moves to
  20260912120000; a deployment that pins that value in its
  own values file (manual Helm installs) must bump it in the
  same upgrade.