Statistics page in WebUI and through an API

[SuperPat45]

I would love to have a full statistics page in the WebUI showing the current activity of the server (by seconds):

• number of call for each endpoint
• number of failed request
• number token IN/OUT
• CPU load
• RAM usage
• GPU load
• VRAM usage
• ...

Also, these data can be represented in charts by days like the OpenAI usage page: https://platform.openai.com/usage/activity

These data could also be available to a new /stats API endpoint.

[mudler]

I'd love to have this in too :+1:

[dave-gray101]

We definitely need to further enhance what we're tracking, but the infrastructure we already have is open telemetry based - so before we spend time working on our own UI, we should probably run some searches for permissively licensed otel compatible display code. At the end of the day something like grafana is always going to be prettier than what we integrate, so I'm assuming that this page should be kept pretty simple