Hello, I’m using the puppet_operational_dashboards...
# forge-modules
r
Hello, I’m using the puppet_operational_dashboards module. Grafana, Influx, and Telegraf are all working fine. All dashboards have data and function correctly, except for the system dashboard (v1 and v2), which doesn’t display any data. Can you please help me identify what I might be missing? My primary/compilers have these classes included. I have a separate host where I run the grafana stack.
Copy code
- puppet_operational_dashboards::enterprise_infrastructure
  - puppet_operational_dashboards::telegraf::agent
s
If I have this correctly memorized, the system metrics are collected by the github.com/puppetlabs/puppetlabs-puppet_metrics_collector module.
And therefor are not collected by telegraf. actuelly it would be nice if telegraf would collect all the data provided by the puppet-metrics-collector module itself. eg. using execs and scrips to source them
r
hmm, does it? it isnt documented than at all 😞 side note: sysstat is installed on the systems
s
can you please give me a cat of all the inputs telegraf has configured right now? I checked the telegraf config.pp and did not see a code snippet connected to system metrics.
r
the telegraf systems dashboard is working and has all data
s
Yes, that was what I understood too
r
okay, mom
i think its some kind of default.
s
Yes, and the dashboard above is not looking for these metrics in the system section. You can just have a look at the query in your working telegraf dashboard. Should look different. I have stumbled over this multiple times in the past. Nice solution from my perspective would be to define the telegraf inputs in the operational_dashboards module and edit the dashboards to use these. Then its a self contained feature
r
¯\_(ツ)_/¯ ja, just want to know how this is supposed to work or if it is b0rken by design and i shall raise a support ticket. just wanna use it 😄 just playing customer here 😄
s
The “send to metrics server” feature in the metrics_collector is marked as deprecated anyhow: github.com/puppetlabs/puppetlabs-puppet_metrics_collector/blob/…/init.pp#…
support tickets are like the nike slogan: just do it 😄
r
okidok 😄
a
This was only intended to work by scraping
sysstat
and
pidstat
metrics from support scripts via one of the plans in the module. It was a mistake to not make this more obvious, and at some point there was a parameter added to not include the system metrics dashboard by default.
❤️ 1