Aligning the uid's for datasources across multiple instances of Grafana

Grafana Configuration
1

  • What Grafana version and what operating system are you using?
    Grafana v12.0.2 (NOT in a Docker container), RHEL8

  • What are you trying to achieve?
    find the best practice for setting up connections to the data sources across more than once instance of Grafana.

We have 3 separate environments: Development, PreLive and Production.
We have an instance of Grafana installed in each.

When we copy a dashboard from one instance to another, the JSON specifies a UID for that datasource in each panel. As the datasource UIDs don’t match in each environment, I’m having to do a find/replace before importing the JSON.

I saw a post that mentions setting up a dashboard variable for the datasource instead.
I can see how to do this but I’m assuming that when i come to copy over Alert rules, I’m going to run into the same sort of issue.

Is there a way go into the Grafana backend somehow and change the UIDs so I can make them the same in all 3 environments?
Does anyone have any better suggestions?
What’s the best practice here?

Thanks for any help you can offer!

All the best :slight_smile:

2

Are these dev/qa/production type of environments?

if you want to keep those uids the same across environments, it probably is best to use provisioning.

3

ooh. That looks interesting. I wasn’t aware of that facility.
So if I define my datasources in a yaml, Grafana will pick them up and use them.
I take it I can set the UIDs for them there?
As I’ve already got Grafana instances with the datasources in place, is there a way to export the settings ready for the yaml file?
Failing that, could someone point me in the direction of a page explaining how I define all the parameters?
I’ve got sources I’ll need to replicate for InfluxDB (v2), Loki & Infinity.
Thanks all!

1 Like
4

but again. you have not answered what all this instances are that requires copying dashboard jsons :wink:

another option is to use git sync. so your dashboards are artifacts committed to gitlab

5

Sorry, I thought that was implicit in my original description. Yes. They are.
All the best.
M.

6

Just so the followup question doesn’t get lost, here it is again:

Thanks again!

7

Here you go

8

Ta.
Does that mean there isn’t a way to export the current settings for the data sources that were entered via the UI?

9

Hi All,
In case it’s of any use to others, I’ve worked out how to get the current settings for the DataSources.

For background - our Grafana runs on RHEL8 and we’ve set it up for https.

  1. Log onto the server
  2. Run the following command to return a list of the data sources from the API:
curl -k https://GrafanaAdminUsername:GrafanaAdminPassword@localhost:3000/api/datasources

Hint: if your password has any special characters in it, you can escape them with a backslash (\).

This should respond with a chunk of JSON with the details you need :slight_smile:

1 Like