Druid configurable count column

[rjdp]

• add "count_column" to KPI model 80fd1d0
• add "Update KPI" query to migration script for upgrade instance 80fd1d0
• read "count_column" from configured value in KPI definition instead of using hardcoded "count" value a318e31
• add KPI validation to make sure selected count_column is of integer type 40856b2
• Fixes for the failing pytests because of this change c61ea1e

[gitpod-io]

[netlify]

✅ Deploy Preview for frontend-sb canceled.

Name	Link
🔨 Latest commit	c61ea1e
🔍 Latest deploy log	https://app.netlify.com/sites/frontend-sb/deploys/62a1a9edbe02670009696479

[manassolanki]

LGTM, Just one question, the count column should be ideally visible in case of druid (or similar data source in future). How are we planning to do the same?

[rjdp]

@manassolanki currently its being used only incase when Druid is the datasource ( in KPI add form as well it comes up only when Druid is chosen) , but we plan to use same field when we handle preaggregation usecases across more DBs/DWs

[manassolanki]

@manassolanki currently its being used only incase when Druid is the datasource ( in KPI add form as well it comes up only when Druid is chosen) , but we plan to use same field when we handle preaggregation usecases across more DBs/DWs

ok, you mean this is hardcode on the frontend side right now.

[rjdp]

@manassolanki it was hardcoded on backend

chaos_genius/chaos_genius/core/anomaly/controller.py

Line 89 in ee1f68b

self._preaggregated_count_col = "count"