Dashboards would be more useful if they could aggregate across instances

[gaggle]

First off, kudos on the fantastic dashboards! They're very insightful.

We've observed that the dashboards only presents data per "Application Instance". This setup prevents us from having an overview of metrics across all instances. Is this a design decision or an unavoidable constraint?

Describe the solution you would like to see
It would be tremendously helpful to see aggregated metrics spanning all instances.

How would you expect this feature to work
Maybe the "Application Instance" filter could accept a wildcard ("*") to display metrics from all instances? Or could a toggle switch between individual and aggregated views?

We were expecting the dashboards to display a comprehensive view by default, and then allowing us to drill down further when/if needed.

Additional context
We deploy to Kubernetes pods so the "Application Instance" dropdown fill up with dozens of entries, and each individual instance lives for very short time as new instances appear every time we deploy.

[amozoss]

I believe it's as simple as setting showAll to true in all the dashboard templates

https://github.com/akoutmos/prom_ex/blob/master/priv/phoenix_live_view.json.eex#L1320
Then have a default value of
"allValue": ".*"

And then you'd want to change instance=\"$instance\" to instance=~\"$instance\" so it can match a regex