Update grafana-dashboard.json

[itay-grudev]

New Cluster Overview

A new cluster overview section at the top that features:

• Base backup and last archived WAL information
• Total CPU and memory usage across all nodes
• An Alerts section
• Last failover/switchover time
• Cluster PostgreSQL version
• Transactions per second
• Volume usage and total database size
• Replication/Write/Flush/Replay lag

Server Health

• Added instance zone
• Displaying the full PostgreSQL version
• Bug Fix: When there are redundant kube-state-metrics instances there are duplicate status gauges
• Bug Fix: Max Connections not displaying a progress bar in the gauge due to missing min and max values.

Before:

After:

Configuration

The configuration parameters are transposed such that every row now contains a parameter, while every column contains the parameter setting across individual database instances. This makes it much easier to scroll through settings.

Before:

After:

New Storage & IO Space and Inode Usage metrics

Added volume space and inode usage gauges.

Accumulated Tuple/IO graph to make the data more comprehensible

Before:

After:

General

• Fixed tooltips and set shared crosshairs of all graphs.
• Bug Fix: Cluster variable query picking up any metric ending with cluster

[github-actions]

❗ By default, the pull request is configured to backport to all release branches.

• To stop backporting this pr, remove the label: backport-requested ◀️ or add the label 'do not backport'
• To stop backporting this pr to a certain release branch, remove the specific branch label: release-x.y

[jsilvela]

Sorry for the delay. Been deploying this locally.
There is a problem with the datasource references as you have them in your files.

"datasource": {
        "type": "prometheus",
        "uid": "${DS_PROMETHEUS}"
}

These will not work properly with the Grafana as we deploy from the quickstart, using the kube-prometheus-stack.
We can substitute the "${DS_PROMETHEUS}" with "prometheus", as that uid is generated for the data source in the kube-prometheus-stack
deployment.

[jsilvela]

Since the JSON file you generated, in particular, you took from a grafana deployed with a different mechanism, I think there is value in keeping it as-is.
The YAML though, I would like to alter to make sure beginners following the quickstart will get the dashboard working out of the box.
So, I plan to add a bit more context in the documentation, and perhaps fix your files but add your JSON untouched but with a new name.

[jsilvela]

On the dashboard itself, the amount of work and polish is ... 👏 bravo.

[jsilvela]

With Itay's latest commit, the dashboard works out of the box if following the quickstart document.
Doing a detail review of the panels...

[sxd]

@itay-grudev fallowing @jsilvela advice, I think we should add the disclaimer into the quickstart with this PR adding that some features may fails because they need a really new/early version of Grafana.

I'm also adding my self as a reviewer to this PR because I really want to try ti! :D

[benoitschipper]

@sxd and @itay-grudev ill happily test/review what I can. Already have the dashboard running :)

[itay-grudev]

@sxd I discussed offline with @jsilvela that the alpha metrics are enabled by default on Prometheus. I think it's a limitation of KinD that the volume data is not reported. I've tested it on AWS EKS and Digital Ocean and there it works fine.

I still need to add the change @benoitschipper proposed.

[sxd]

@itay-grudev I've added a warning message that will help users to understand why the graphs doesn't have data, can you review it please! I understand that those are limited in KinD, but lots of users use KinD for testing and at least we need to be nice and put the warning message

[benoitschipper]

@sxd I discussed offline with @jsilvela that the alpha metrics are enabled by default on Prometheus. I think it's a limitation of KinD that the volume data is not reported. I've tested it on AWS EKS and Digital Ocean and there it works fine.

I still need to add the change @benoitschipper proposed.

Ill check/review it by testing it as soon as you have added the regex proposal so I can confirm if it works.

[itay-grudev]

@benoitschipper Done. Test if it works for you now.

[benoitschipper]

@benoitschipper Done. Test if it works for you now.

Will do!

[benoitschipper]

@itay-grudev I loaded the Dashboard and went through all the graphs on a live cluster in two seperate clusters. All worked without me making any changes (only the name). Works like a charm!