Runbook
Procedure: Restart the payments worker
Service: Customer payments API
Owner: Platform team
Last reviewed: 3 August 2030
1. When to use this runbook
Use when payments are queued for more than 10 minutes and the dashboard shows the worker as down.
2. Access and tools you need
Read access to the monitoring dashboard
Shell access to the production jump host
Member of the on-call group
3. Checks before you start
1. Confirm the alert is real: open the Payments dashboard and check the queue size.
2. Check the status page for a provider outage.
3. Tell the incident channel you are starting.
4. Steps
1. Log in to the jump host.
2. Run the service restart command for the worker.
3. Wait 2 minutes.
4. Check that the queue size is falling.
5. Check three recent payments completed.
5. How to confirm it worked
Queue under 50 and falling, no new errors for 10 minutes.
6. If it does not work
Stop the worker, switch the service to the previous release, and escalate to the payments engineer.
7. Who to call next
| Who | How to reach them |
|---|---|
| Payments engineer on call | via the paging tool |
| Platform lead | +00 000 000 0000 |
| Customer support lead, if customers are affected | support channel |
8. Warnings
Never restart both workers at once. Do not clear the queue by hand.
9. Afterwards
Write a note in the incident log with times and what was done. Raise a ticket if the cause is not known.