I'm building a OneAgent plugin in order to create alerts for container level events, since Dynatrace can't do this (yet). The plugin gets a list of containers and stats for each container from the docker stats API.
I'm uncertain about the correct implementation of process snapshots, specifically the correct combination of pgi_name/id in the plugin code and processTypeName in the plugin.json schema.
My current configuration is pgi_name('dockerd') and "processTypeNames": ["DOCKERDEAMON"], but this is not working for some reason. The plugin builds successfully without any errors.
Any insights would be appreciated.
As far as I remember (been a while since I've played with the OneAgent SDK), but I remember it this way:
The "processTypeNames" or "processTypes" determine if plugin should be "executed". This means if process of give type specified in the array is found, the plugin will get executed. On the other hand, the pgi_name/id in the plugin code determines for which processgroup will the data be sent. AFAIK should be able to send out metrics for process groups of any types (not just the ones defined in plugin.json).
Did you have a look how the ruxit.python.docker plugin is written? The Docker monitoring is also performed using the python plugin.