Batch API Requests expiring without any progress at all

@gokulraya - is this issue potentially back? Our system sends about ~10 batch jobs every night, the last three days I have seen most of the jobs complete within 2 hours, but 2 batches have gone 23 hours+. This morning I see that 6/10 have made 0% progress.

It seems similar to this issue - Gpt-4o Batch Processing Jobs Response Time Increased Significantly, Causing Job Timeouts/failures
or this issue - Run status stucked in 'queued'