Skip to main content

Notification retries

When Seiri sends an alert to Slack, Discord, email, or another channel and the channel call fails, Seiri retries with backoff so a transient outage does not drop the page.

This is alert delivery retry — not retrying your cron job, and not Seiri calling your application webhook.

Practical tips​

  • Prefer stable channel webhooks / bot tokens.
  • Test channels when attaching them to a monitor.
  • Keep alert noise down with sensible grace periods so retries are not amplified by flapping.

Job-side pattern: always run the real work first; only then ping Seiri. Swallow ping errors.

See Backoff and Integrate a job.