-
Notifications
You must be signed in to change notification settings - Fork 504
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
retries exceeds totalTimeoutMs request budget #2369
Comments
Hi @thekia19! I wasn't able to reproduce this issue with a simplified version of your config. Here's what I used:
I then launched a simple webserver which always responds with a status code 500 and sent a request to it through Linkerd using curl. In the webserver logs I see that there were a total of 5 requests (the original plus 4 retries) before the timeout is reached, after which no more requests are made and Linkerd returns a response with the message to curl:
If you can tell me more about the exact behavior of your server that causes the behavior you describe, that would help me to reproduce the issue. |
Closing due to inactivity. |
Filing a Linkerd issue
Issue Type:
What happened:
Request had been retrying for 5 hours whilst request budget (totalTimeoutMs) was set on amount 5s
What you expected to happen:
Request is failed when budget is exceeded
How to reproduce it (as minimally and precisely as possible):
Sorry. I can't provide steps to reproduce. But I provide config causing the issue and config eliminating it.
Environment:
Config causing issue
Config fixing issue (only 2 parameters differ)
When maxMs (set on 10000) was reached request was retrying again and again for 5 hours each 10 seconds causing an exception in linkerd (trace is below). On the same time application 10.20.93.22:6002 successfully responded to linkerd, but linkerd continued retrying request.
It was happening up to 20 times per day whilst application send 100k rpm to linkerd meaning that it's not critical but unpleasant thing to experience.
Is it an expected behavior if jittered backoff maxMs has higher value than request budget?
I believe it shouldn't be a such case. Please correct me if I wrong.
Many thanks
The text was updated successfully, but these errors were encountered: