Add 'store' label to metric pd_cluster_status.

[SerjKol80]

Enhancement Task

Currently gauge metric pd_cluster_status is aggregated across all stores. This is pretty inconvenient, because if you have long running deployment across TiKV component this metric indicate problematic store for a long period of time and it will not be possible to distinguish different tikv stores contributed to this metric.

What is suggested. Instead of emitting counts across all stores here, we will emit the same gauge metrics with additional label 'store' with values either 0 or 1 in observe() method. This will allow to get more detailed view from the metric. And aggregation across all values of 'store' label will let you to get the same metrics as today.

[ti-chi-bot[bot]]

Welcome @SerjKol80! It looks like this is your first issue to tikv/pd 🎉

[bufferflies]

Enhancement Task

Currently gauge metric pd_cluster_status is aggregated across all stores. This is pretty inconvenient, because if you have long running deployment across TiKV component this metric indicate problematic store for a long period of time and it will not be possible to distinguish different tikv stores contributed to this metric.

What is suggested. Instead of emitting counts across all stores here, we will emit the same gauge metrics with additional label 'store' with values either 0 or 1 in observe() method. This will allow to get more detailed view from the metric. And aggregation across all values of 'store' label will let you to get the same metrics as today.

you can find the store label details in dashboard