This repository has been archived by the owner on Feb 18, 2021. It is now read-only.
Error from send: Message too long #71
Comments
|
This happens when you try to send metrics to a UDP backend that is larger than your network interface's MTU (probably somewhere around 1500 bytes). You can either use shorter metric keys or switch to TCP for the backend to resolve this problem. |
|
Dear @JeremyGrosser , |
theatrus
referenced
this issue
in lyft/statsrelay
Oct 29, 2017
Previously, the termination hooks would tear down resources (including destroying server collections) while still running from within the ev_loop. ev_break is only guaranteed to exit the active loop after all pending signals in that invocation have been processed. This led to a race and a (annoying but mostly harmless) crash. The loop exit code after ev_run() already does all the teardown, so we are re-using that teardown logic and only stop accepting new connections from the signal handler (not that they would be processed most likely).
theatrus
referenced
this issue
in lyft/statsrelay
May 1, 2020
Previously, the termination hooks would tear down resources (including destroying server collections) while still running from within the ev_loop. ev_break is only guaranteed to exit the active loop after all pending signals in that invocation have been processed. This led to a race and a (annoying but mostly harmless) crash. The loop exit code after ev_run() already does all the teardown, so we are re-using that teardown logic and only stop accepting new connections from the signal handler (not that they would be processed most likely).
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Hello All,
I start statsrelay instances and use statsd-tg to do performance testing. After push metrics minutes, statsrelay show error "Error from send: Message too long". Is it because of the queue full and drop message?
Test script
statsrelay1.conf
ERROR
The text was updated successfully, but these errors were encountered: