Simulation

Resilient job runner

Run a 240-record extraction job, throw failures at it, kill it halfway, and watch it resume from its checkpoint.

Interactive demo

Simulation. No real requests are made. The runner models how my production jobs behave: timeouts, rate limits, permanent errors, crashes and resuming from a checkpoint. Outcomes are random, so every run is different.

Job · export 240 records

idle
0done
0retries
0failed
0.0selapsed

Conditions


done working retrying failed

Log

checkpoint: none
What this shows

Retries with backoff, rate limits, checkpoints, resuming without duplicates, and honest final reports.

Long data jobs fail. A portal times out, an API asks you to slow down, a server restarts at 3 a.m. What matters is what happens next. This runner shows the habits I build into every long job:

  • Retries with backoff for temporary errors, with a limit, so a dead record doesn't loop forever.
  • A shared pause when the source says "slow down", instead of every worker hammering it.
  • Permanent errors recorded, not retried, and listed in the final report.
  • A checkpoint after every record, so a crash costs seconds and finished records are skipped on resume.
  • Idempotent writes keyed by record ID. A crash can land between writing a record and saving the checkpoint; an upsert means that record is rewritten, not duplicated. Without it, "resume" quietly creates duplicates.

Everything here is simulated in your browser: no requests are sent, and the error rates are yours to set.

Close to: permit exporter, mobile proxy farm.

Related service: Infrastructure, reliability and rescue

Next step

Need this for real?

The demos run on sample data. Tell me about your source, your workflow or your data, and I will tell you what a real version would involve.