You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We noticed while looking at client logs, that requests that are setup with 30s or 1mn timeouts can sometimes not abort/hangup once the timeout is reached. We've seen requests last up to ~10mn.
The consequences are that under heavy load and slow response times from a remote host, our server is not freeing up request threads as fast as we thought, and performance is degrading.
The text was updated successfully, but these errors were encountered:
amcclain
changed the title
fix http client timeout so that requests really do abort upon timeout
Fix http client timeout so that requests really do abort upon timeout
May 10, 2023
amcclain
changed the title
Fix http client timeout so that requests really do abort upon timeout
Fix JSONClient timeout so that requests actually abort on timeout
May 10, 2023
We noticed while looking at client logs, that requests that are setup with 30s or 1mn timeouts can sometimes not abort/hangup once the timeout is reached. We've seen requests last up to ~10mn.
As explained in this post
https://www.baeldung.com/httpclient-timeout
our current timeout configs do not hard abort a request.
The consequences are that under heavy load and slow response times from a remote host, our server is not freeing up request threads as fast as we thought, and performance is degrading.
To get real request aborts, we want to add a
TimerTask
with amethod.abort()
call to themethod
passed into https://github.com/xh/hoist-core/blob/develop/src/main/groovy/io/xh/hoist/http/JSONClient.groovy#L145The text was updated successfully, but these errors were encountered: