Hello,
Recently we have been seeing 504 errors on the ELB status logs (AWS) for Jenkins. When investigated, we came to know that, to keep the page on browser in sync with latest information, the browser keeps on polling data using ajaxBuildQueue and ajaxExecutors endpoints. When the logs for these requests and their responses are observed, we saw that intermittently 504 errors are popping up. This tripped our thresholds for 5xx errors. We want to know how to eradicate the 504s as it might also impacts performance if we start to see more of these errors.
When observed, we found 26 of the 504s on the logs over a period of 15 mins. While many of the requests land on ELB one after the other, only one out of 10 subsequent requests to the mentioned endpoints produced the 504s.
Not sure if this is related to performance tuning where in we tune the thresholds in configuration somewhere or is it something to do with jenkins itself. Any suggestions are highly appreciated.
Regards,
Prasad Velidi