On very large PMM systems Prometheus cannot be stopped in 10 seconds on 'docker stop' command.
Docker kills all processes in a container after 10 seconds.
After kill Prometheus database becomes corrupted.
workaround: create pmm-server container with '--stop-timeout 600' option
The question is, where to put this option:
1) in the default configuration string for 'docker run'.
Pros: will not add additional "If" for User
Cons: will make our docker run command more complicated (but, it' an already not easy )
2) Ad this option as optional after default command with Note that if you planning to have large installation use this option to avoid data corruption
Pros: Initial command will remain shorter
Cons: someone can miss our recommendation to larger systems and have a problem with production
TODO: Mention in Documentation this option (--stop-timeout) for installations where planned a big amount of data.