Does anybody know a good way to troubleshoot this ...
# fleet
Does anybody know a good way to troubleshoot this kind of errors?
Copy code
Sep 08 10:27:24 devo-ea-manager fleet[860]: 2021/09/08 10:27:24 http: TLS handshake error from EOF
I am doing some scale testing, and i have an AWS t3.xlarge box with 2k agents where everything seems to work correctly but this error appears periodically, any suggestion to understand why this could be happening?
I have read that this could be related with ulimit, but we have set limits to more than 16k and opened FDs at the time of the error is under 100.
There is a new debug command that might be helpful
Thanks for your answer. Mainly connectivity seems to work but the error appears sporadically for some agent connections (which seem to connect properly the next time they check in). That command seems to check only connection from a global pov which i would say it is working. Do you think i could use it to get more info on what could be going on anyway?
I think i found my issue, when creating many dockers in the same box for testing i was hitting an issue with ARP overflow. Fixed that in the agents server and the issue got fixed.
👍 1