The IO issue is still present after purging the agent, marking sure that all old config files are gone from the vm and rebooting multiple times in between.
The long break in between were a couple of days were I had stopped the agent. Didn’t find time to look at the issue until today.
It seems that after running 2.1.0p17 with enabled TLS the IO issue is even worse.
Currently I am out of ideas.
check_mk_agent runtime of on this server:
real 0m2.701s
user 0m1.856s
sys 0m0.266s
Edit: We have the APT plugin usually running. I removed that to see if that might be the cause.
real 0m0.608s
user 0m0.253s
sys 0m0.056s
Pulling the APT plugin once an hour with the async config might help here. It doesn’t really have to run every time the agent is pulled.