Hello,
I am creating a “fork” of the Grafana agent snap to have a more recent version, as well as manage a new mode for the agent, the flow mode.
Snap name: rob-cos-grafana-agent
Account name: Ubuntu Robotics Community
The current revision uploaded requires two manual reviews.
etc-grafana-agent, system-files:
Necessary to read /etc/grafana-agent.river configuration file from the host.
proc-sys-kernel-random, sytem-files:
Necessary to read /proc/sys/kernel/random/write_wakeup_threshold, /proc/sys/kernel/random/read_wakeup_threshold, /proc/sys/kernel/random/poolsize and /proc/sys/kernel/random/urandom_min_reseed_secs so the Grafana agent node-exporter can collect information about the CPU.
In general this seems reasonable but perhaps it would be worth collaborating with the existing publisher of the grafana-agent snap to add your changes there.
Also for consistency, should the etc-grafana-agent plug be named etc-grafana-agent-river?
Hello,
I have already opened tickets on the official grafana-agent snap repository and got the confirmation that some won’t get done in the near future.
I have added the features in my fork in a temporary way (only supporting my use case).
This snap is meant to stay unlisted and is meant to be used in the context of a demo.
Integrating my changes in the official repo might take months.
Does that make sens to change the etc-grafana-agent into etc-grafana-agent-river since the previous plug didn’t have the -yaml at the end while it was reading a YAML. river is simply the configuration file extension.
Given the nature of the snap and due to the delay in merging these changes into the official repo, +1 from me for connecting these system-files interfaces. what do other @reviewers think?
@RoboticsCommunity - just to confirm, is connect or auto-connect desired?