Cloud Instance Congestion: Thousands of executions stuck in queue after high-traffic event

[Prafful_Agrawal]

The Issue: Yesterday, a campaign sent approximately 60 contacts per minute to our n8n Cloud instance. This briefly exceeded our concurrency limit of 50, resulting in 2,000–4,000 executions getting stuck in the queue.

Current Status: Even after manually stopping the old executions, the instance is behaving randomly. New, unrelated executions are getting added to the queue and not executing, even though traffic has returned to normal levels. We suspect backend congestion or “zombie” worker processes.

Questions for the team:

1. Is there a way for support to clear the backend queue or “stuck” workers for this instance?

2. Is a manual restart of the cloud instance feasible or recommended to resolve this?

3. Are there any known issues with queueing on Cloud after a concurrency spike?

[Mohammed_Rifad]

I face issue with my 3 self hosted n8n, I see 502 error. Its hosted in railway and render. Both failing, not sure whats wrong.

[Parintele_Damaskin]

Hey @Prafful_Agrawal !

Welcome to n8n community!

I would suggest to restart your workspace from within Admin Panel, since you are on Cloud version. Access to back-end is available only to the n8n team.

If after that you see some “ghosts” there, is recommended to report at [email protected] , and with the ticket number and workspace ID, send DM privately to n8n member(they have the badge attached to avatar), if they are around here.

Cheers!

P.S @Mohammed_Rifad , please open a new topic with your problem.

Cheers!