Skip to main content
Scheduled-job monitoring reverses the HTTP-monitoring flow: your job reports its lifecycle to Still200. Still200 already knows when the job should run from its cron schedule, so a missing start ping becomes a missed run. For Python jobs, the Still200 SDK wraps this lifecycle in a decorator. Jobs in any language can send the same signals through the manual HTTP integration on that page. Python is the first supported SDK language, with more language SDKs planned. Still200 detects three failure modes:
  • Missed: no start ping arrived by the scheduled time plus its grace period.
  • Failed: the job called the fail endpoint.
  • Timed out: the job started but sent no terminal ping before its configured maximum runtime.
The same model works for AI agents that run on a schedule. Treat each scheduled invocation of the agent as one job run, and Still200 will catch an agent that never starts, errors out, or gets stuck looping past its maximum runtime. See the AI agent loop example.

Configure a schedule

Integrate your job