micrometer-metrics/micrometer

Statsd UDP epoll warn logs are verbose when statsd down

Open

#2,624 opened on May 27, 2021

View on GitHub
 (9 comments) (9 reactions) (0 assignees)Java (935 forks)batch import
help wantedregistry: statsdtype: task

Repository metrics

Stars
 (4,220 stars)
PR merge metrics
 (Avg merge 1d 21h) (100 merged PRs in 30d)

Description

Testing on Linux, I can see we get logs like the following with #2611 (and therefore epoll enabled) when the statsd daemon isn't running:

2021-05-28 03:11:45.544  WARN 6373 --- [    udp-epoll-1] i.m.s.reactor.netty.channel.FluxReceive  : [id:35f314ec, L:/127.0.0.1:60434 - R:localhost/127.0.0.1:8125] An exception has been observed post termination, use DEBUG level to see the full stack: java.net.PortUnreachableException: readAddress(..) failed: Connection refused
2021-05-28 03:11:45.554  WARN 6373 --- [    udp-epoll-1] i.m.s.reactor.netty.channel.FluxReceive  : [id:35f314ec, L:/127.0.0.1:60434 - R:localhost/127.0.0.1:8125] An exception has been observed post termination, use DEBUG level to see the full stack: java.net.PortUnreachableException: readAddress(..) failed: Connection refused

This is a bit verbose compared to without epoll enabled, in this scenario of the statsd daemon being down. With epoll enabled and the statsd daemon running, metrics are received and no logs are output, as expected.

Originally posted by @shakuzen in https://github.com/micrometer-metrics/micrometer/pull/2611#issuecomment-849839027

Contributor guide