Skip to content

Worker Runtime Operations

Typical operational checks

  • confirm that the tenant is enabled
  • inspect whether the relevant jobs.*.config.enabled flags are true
  • compare current time against configured repeat_interval, repeat_interval_cold, plan_run, or deep-sync settings
  • check whether a missing UI update is actually caused by an upstream dependency still waiting on packages

Common runtime patterns

Situation Likely explanation
packages visible, payloads stale payload loop blocked, delayed, or not enabled
messages stale, alerts stale package dependency not fulfilled or message loop delayed
alerts stale but messages current alert loop delayed, disabled, or lock-blocked
archive missing plan_run not reached yet or archive job disabled

Important scheduling examples

  • archive loop uses a planned daily time such as 02:00
  • packages loop may use repeat_interval=600
  • payload loop may use repeat_interval=30
  • alerts loop may use repeat_interval=60
  • messages can combine fast and slower intervals through hot and cold settings

These are examples from current code paths and defaults, not hard platform invariants.

Why worker/runtime is a frequent root cause

Most "data is missing in the UI" incidents start in one of these places:

  • a loop is disabled
  • a lock was not acquired
  • an upstream dependency has not completed
  • runtime timestamps still place the next run in the future