bradh352 commented on issue #12574:
URL: https://github.com/apache/cloudstack/issues/12574#issuecomment-3850709002

   @chunkyen nothing here is k8s specific at all, it would be a problem for 
anything that couldn't handle such short timeouts.
   
   I do agree there should be a per-service tunable to set time timeout.  
However, the default values are so low that a global tunable seems reasonable 
at this point in time.  It looks like this timeout could even trigger for a 
query taking a long time before it started returning data since it is unaware 
of the underlying protocol when it is raw tcp and not http.  I think for 
non-http flows where haproxy has no knowledge of the underlying communication 
this is really really bad.  That said, I want to maintain as much compatibility 
as possible with existing deployments.
   
   It would be great if the software side could handle such timeouts, but 
something that short may not be feasible for long-running  queries.
   
   We can definitely discuss taking this up on a per-pool basis though, kind of 
like configuring sticky-ness is per-pool.  I just don't know what would be 
acceptable upstream since I'm fairly new to the community.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to