Open 10101010101010001 opened 3 years ago
Hmm that's doable maybe. But lot of metrics are gathered from sysfs cgroups - and these can't be scraped remotely (imo). Didn't looked much to new api in 4.x
lxd, so maybe it can be done. Will try.
Purpose of this is to have telegraf & this script installed on all lxd hosts and single influx/grafana targe anywhere. I have influx&grafana in one of my lxd containers.
But I'll try to do something with this - don't know when.
@wociscz I was able to test this script with a remote host by modifying your script:
Here's the added code:
import urllib3
urllib3.disable_warnings()
if __name__ == '__main__':
from pylxd import Client
client = Client(endpoint='https://url.domain:8443',cert=('/home/user/clients/host/client.crt','/home/user/clients/host/client.key'), verify=False)
Note: Some of my remote hosts are running Ubuntu 16.04/18.04/20.04, some are running on Centos 7/8.
Error: output.append(header + ",metric=status running=" + str(metric['running']) + ",processes=" + str(metric['processes']) + ",cpuprio=" + str(metric['cpuprio'])) KeyError: 'cpuprio'
Thanks for your work, I appreciate your efforts. I'm still learning python or I'd implement this myself but could you add the ability to gather the statistics of remote hosts attached to lxd?