a month ago
Hi Railway team,
We're on the Pro plan (project nurturing-quietude, ID 881b5d3d-990c-46aa-87b8-3018c4751f4e, service coco, ID 8bbc67cc-00a3-4749-9792-3ab11162ed7f, environment c9312f85-bf1f-4035-982a-66a0ad564b49). The coco service runs parallel coding-agent jobs, and each job spawns many short-lived processes and threads. The 1,000 PID limit per container is now our bottleneck: CPU and RAM are far from saturated, but we can only run about 5 jobs in parallel instead of the 15 we're aiming for.
Could you raise pids.max for this service to 4000 (or 2000 if that's the most you can do)? We already cap thread pools per job and reap processes. Happy to give more details.
Thanks!
1 Reply
Status changed to Awaiting Railway Response Railway • about 1 month ago
a month ago
1k PIDs is the most we can do on the Pro plan, instead, please horizontally scale out with replicas, as each replica will get it's own 1k PID limit.
Status changed to Awaiting User Response Railway • about 1 month ago
a month ago
This thread has been marked as solved automatically due to a lack of recent activity. Please re-open this thread or create a new one if you require further assistance. Thank you!
Status changed to Solved Railway • about 1 month ago