search

ytx1991 / ChiaMonitor

An approach for monitoring multiple Chia harvester in one web app based dashboard
150 stars 41 forks source link

Looks incredible, but, not working at all #5

Closed hyacin75 closed 3 years ago

hyacin75 commented 3 years ago

I'm not sure if there are major problems here, or issues with the instructions, or if I'm just missing something?

I've got Prometheus and Grafana up, no problem, was already running Grafana, just had to spin up a Prometheus instance.

It looks like it is expecting a debug.log and a farmer_summary.log? But neither of these exist with a default setup. The only logging config I see in the Chia configuration pertains to running a full node, but I'm deploying this on my harvesters... so... not sure how to get them making a debug.log. Also not sure about this farmer_summary.log at all ... I tried firing up the first command from the run script in a separate screen session and redirecting it out to that file as someone mentioned in another issue here, then running mtail by itself, as, the way the start script is written, it just hangs up on the first command and never runs mtail ... that got mtail listening on on 3903 at least, but still no data on my dashboard.

You say Prometheus and Grafana need to be running on the full node? I'm, not sure why, at all? I see nothing in the config or setup instructions for them to pull any information from the full node ...

Totally lost here, and shocked to see comments in other issues saying "Got it working" as that almost looks impossible from the pieces here.

Any advice? Or plans to add a little more to the instructions and clean up big obvious issues like the fact that the startup script is fundamentally broken?

Thanks!

hyacin75 commented 3 years ago

Ok, so, more 'notes in issues' for the next person to come along looking for more info -

The datasource HAS to be called "Chia", and not only that, it HAS to be the default datasource ... that's a couple hours of my life I'd like to have back.

Remote harvesters WILL write logs, I guess they just don't have much to say at the default WARNING level. You have to update the config so the loglevel is INFO and then restart the harvester.

Best hack I can come up with if your system does in fact have watch is to just put an & at the end of that first line in the run script. If the system doesn't, you can write a small script or whatever to run chia farm summary and write it out to ~/.chia/mainnet/logs/farmer_summary.log

I think that's all I've got so far ... graphs are finally showing something ... need to enable the rest of my nodes again now.