feat: usage of predict_linear function in volume dashboard

[Hardikl]

[Hardikl]

Update from the investigation so far,

As discussed, I have tried to use the predict_linear function this way, Sample interval: 3 days, Prediction interval: 15 days.

There are few challenges while plotting this graph in Grafana

• Query seem to be too slower as predict linear function takes topk query as input which is also slower in nature due to range and avg_over_time.
• While plotting the graph, If we use sample interval as 3d then we can't provide the Time shift for more than 3d.
  Time shift is the value which can shift whole graph in right(future) or left(past) direction.

My idea was to plot a graph this way:
from current_date-15d(past) to current_date+15d(future) and current date will be at center. In this case, Time shift should be 15d.

[Hardikl]

[Hardikl]

[Hardikl]

Instead of using physical used and logical used, now using size used with clamping at 0% and 100% and made table size as full width which shows node and cluster data as well.

[rahulguptajss]

Need Datacenter column in table.
Cluster should be a column after datacenter (at start) in table.
Default sorting should be on predicted value.
Is gradient bar percentage display better than text percentage value?
Node can be removed from table.

[Hardikl]

• Datacenter column added
• Cluster column moved to 2nd
• Node column removed
• % changed to gradient
• Default sort desc for predicted value.