Webb Brown

@AjayTripathy friendly ping...

Would love to see first party support here! @skonto any suggested contacts on the IBM team? cc @alexkubecost @trentbroderick

Hi Omer, good question. What exactly are you looking to monitor? If cost model metrics are being collected? Or that the cost model is correctly able to serve data? Or...

The best approach to do this continuously is likely to use the [aggregated cost model API](https://github.com/kubecost/docs/blob/master/allocation-api.md#aggregated-cost-model-api). This reports the same data as the Kubecost frontend and would go above and...

Depends exactly what you are looking to confirm, but you could do a short time window, e.g. 1-5 minutes, like that!

Yep, this can be accomplished! You would need to monitor several metrics for this to be complete, or you could do the simple test that our /metrics is UP.

Depends on your cluster, e.g. using GPUs, etc. But monitoring a subset of these metrics is likely best: https://github.com/kubecost/cost-model/blob/master/PROMETHEUS.md#available-metrics

Yeah, we can explore adding something like this for you if it would be helpful and our product is overall a fit for you!

@smitthakkar96 any other checks in diagnostics.html that would be most helpful? All of the check below currently have underlying Prometheus measures that you can monitor. I do agree adding one...

This is great, thanks for sharing @MrColeC! We have this on the docket for our upcoming release (1.94). We'll start to review next week and will find owners soon after....