Long-running Serverless Requests on Runpod Execute Twice, Doubling Billing Costs
I’m experiencing an unexpected behavior when running long requests (with data payloads around 200k samples) (Clustering using berTopic). After the job finishes and posts the results, it unexpectedly triggers another execution of the same request. This behavior causes the job to run twice, effectively doubling my costs. Notably, when I run similar jobs with smaller payloads (e.g., 10k or 20k samples), they execute normally without any duplicate runs.
4 Replies
Hey @Talion thanks for sharing this issue. Could you share a request Id so we can look into this for you?
b0f0995d-9714-4586-a77a-ae2f3cac391c-e1
this request timed out after 6 minutes not even executed but the regular behaviour is that the long requests executed twice
{
"delayTime": 456267,
"executionTime": -1147,
"id": "f73b3f63-558a-46ac-a882-525935ad2c4a-e1",
"output": {
"job_id": "f73b3f63-558a-46ac-a882-525935ad2c4a-e1",
"request_id": "string",
"status": "success"
},
"retries": 1,
"status": "IN_PROGRESS",
"workerId": "pl24d3mls3ewcx"
}
this request executed successfuly and then executed again
Thanks i'll take a look and get back to you!
@Talion we found the issue in our logs, thanks again for flagging it. We’ve credited you for the time lost and will prioritize the fix.
@Gabe Thanks, Gabe! I appreciate your help