Question about the Difference Between Copilot Metrics Projects - copilot-metrics-viewer and copilot-metrics-dashboard

[flavio-araujo]

Hi folks,

I'm curious about the difference between these two projects for viewing GitHub Copilot metrics:

• https://github.com/github-copilot-resources/copilot-metrics-viewer
• https://github.com/microsoft/copilot-metrics-dashboard

Considering that GitHub and Microsoft are now one large company, I wonder why there are two separate efforts. Has there been any discussion about potentially joining forces to create a single, more powerful platform that could benefit all users?

Any insights into the goals, features, or target audiences of each project would be greatly appreciated.

Thanks!

[karpikpl]

that's a good question.
There are some efforts at GitHub to include metrics as part of the product, but that hasn't been the case yet.

Few Microsoft / GitHub employees took this into their own hands and created few solutions.
Due to the range of customers we're supporting, there are different requirements around cost, hosting platforms, networking, authentication etc. Hence different solutions.
There's no single official one
There's also the PowerBI one -> https://appsource.microsoft.com/en-us/product/power-bi/github.github-copilot-metrics
And Kubernetes Grafana Elastic Search -> satomic/copilot-usage-advanced-dashboard

[flavio-araujo]

Hi everyone,

I've had a look at the microsoft/copilot-metrics-dashboard project and I must say, it's quite impressive. The user interface is excellent, and I really appreciate that it already handles data storage in a database and has integrated login functionality. These are crucial features for a robust solution.

However, I have some concerns regarding the technology choices. It seems to be heavily tied to C# and Azure Cosmos DB. While I understand Microsoft's inclination to use its own stack, it would be immensely beneficial if the solution adopted more generic and cloud-agnostic technologies. This would greatly simplify deployment across various cloud providers like AWS, GCP, OCI, etc.

Instead of C#, using languages like TypeScript or Python, along with support for open-source database options such as MariaDB, PostgreSQL, MongoDB, Apache Cassandra, or CouchDB, would make the solution far more flexible and accessible to a wider range of organizations.

Regarding the Power BI project, while it also addresses data storage, it inherently limits its utility to companies that already use Power BI. This isn't a universal reality for all businesses, potentially excluding many who could benefit from historical Copilot metrics.

Ultimately, I believe a unified and more open-source approach would create a truly powerful and widely usable platform for everyone.

What are your thoughts on this?

[karpikpl]

I think you could benefit from satomic/copilot-usage-advanced-dashboard - it's fully open source, no dependencies on Microsoft, runs on kubernetes.

It's hard to find that sweet spot between interoperable app and one that has features users need.
I've been thinking about adding storage to copilot-metrics-dashboard using https://nuxt.com/modules/storage to keep it cross platform.

[flavio-araujo]

Looks like the github team will delivery a native solution: github/roadmap#1132