At random intervals, the MySQL-instance becomes unresponsive, and needs to be restarted with 'kill -9'
The exporters seems to be running, but mysql:* doesn't report any statistics. linux:metrics does, and shows cpu-usage at 0%.
Verified the issue with multiple servers, on multiple MySQL versions (5.6, 5.7 upgraded from 5.6, 5.7 & Percona Server 5.7)
Some servers has been running without any issues for a year, and some only manages a few weeks. No common denominator that we can find. We've been able to reproduce this in a test environment, so we can do as much debug as required.
Attached is output from pt-pmp on the hanging instance