Observability starter kit writeup #573
Conversation
This document provides an overview of the Observability Starter Kit rock-on, including installation instructions, prerequisites, and component descriptions.
Added detailed instructions for setting up the Observability Starter Kit, including user and group creation, share configuration, and Grafana dashboard setup.
Updated the observability starter kit documentation for clarity and consistency, including changes to image widths and text phrasing.
|
@kanecko Just a quick note here: there seems to be a new broken link in these additions: [EDIT] I've now addressed the prior failure of this sort, i.e. an example URL being picked up as a real one and failing our link check. Take a look at the following PR to see who we look to have standardized on formatting for this type of example URL: I.e. we code format then. Above PR as the details of the fix now required in this branch. |
|
@kanecko It loos like we now have a timeout error with one of the new links introcuded in this branch:
Could you take a look at that? Thanks for merging in the existing fix re the zoneminder doc entry by the way. I.e. does this also show up when you run the linkcheck locally? |
|
nope, must be an intermittent problem: |
phillxnet
added
the
needs review
Hooverdan96
left a comment
Hooverdan96
left a comment
There was a problem hiding this comment.
I think this is a great writeup for setup and a tutorial on how to do one's first dashboard! I left a few comments/suggestions below.
| #. Under the *SYSTEM* menu, select the *Users* menu item. | ||
| #. Add two users: osk-grafana and osk-opentelemetry. Assign the group with the same name to each of them. | ||
| #. Under the *STORAGE* menu, select the *Shares* menu item. | ||
| #. Add three shares: osk-opentelemetry-collector, osk-victoria-metrics, osk-grafana |
There was a problem hiding this comment.
I've noticed that the example in the Rockon installation wizard is different compared to what you have here in the documentation: osk-opentelemetry-collector.
Not a big deal, but since the other two are aligned with your example in the Rockon definition, it might make sense to harmonize this one as well (either change it here or in the Rockon definition).
| #. Example*: :code:`chown -R 472:472 /mnt2/osk-grafana` | ||
| #. Example*: :code:`chown -R 10001:10001 /mnt2/osk-opentelemetry-collector` |
There was a problem hiding this comment.
Is changing the ownership setting in the WebUI after creating the share is not sufficient? Because now the UI information and the actual ownership on the disk is different from each other. Not a huge deal, but want to make sure that there was not a specific reason it had to be done via the command line?
| #. Once logged-in, click on **Data sources**. You can find it on the left-side menu, below *Connections*. | ||
| #. Click on the big **Add data source** button. | ||
| #. Find and select "VictoriaMetrics" in the list. | ||
| #. In the HTTP **URL** field, input :code:`http://osk-victoria-metrics:`. |
There was a problem hiding this comment.
| #. In the HTTP **URL** field, input :code:`http://osk-victoria-metrics:`. | |
| #. In the HTTP **URL** field, input :code:`http://osk-victoria-metrics:8428`. |
That's the only way I was able to save & test the data source.
| #. Select the "Code" option. You will find it at the bottom, next to the "Run queries" button. | ||
| #. Replace the current metric with the following expression: ``rate(system.disk.io[$__interval])`` | ||
| #. Change the unit from "Data / bytes (IEC)" to "Data rate / bytes/sec(IEC)" | ||
| #. Scroll up to the "Tooltip" section, and select the "All" option, and then the "Descending" option. |
There was a problem hiding this comment.
| #. Scroll up to the "Tooltip" section, and select the "All" option, and then the "Descending" option. | |
| #. Scroll up to the "Tooltip" section, and select the "All" option, and then the "Descending" option. | |
| #. Click on the "Run queries" button to update the dashboard, if it hasn't automatically already. |
Just to be consistent
| The time window in this case is denoted by ``[$__interval]`` which is bound to the time range | ||
| that you can select above the chart (next to the "Refresh" button). | ||
|
|
||
| The changed tooltip mode has enabled us to mouse-over the chart to see a list of current values of all devices. |
There was a problem hiding this comment.
since this is not a nvme based system, my legend has some awfully long names. Is there a setting that I missed to reduce that to device name only?
I guess, when changing the label in the query options from auto to custom and something like: {{device}}.
Might be worthwhile to add that in as well to reflect the label format you are showing in the screenshots
3 participants
Fixes #572
This pull request's proposal
An installation and first dashboard set-up for the Observability Starter Kit rock-on
Checklist