Proposal

I would like to scale kubernetes workloads based on time-series metric data fetched via OpenSearch query.
A typical use-case would be to activate scaling once average value from a specific time period would exceed a threshold value.

Scaler Source

Averaged time series metrics compared against a given threshold value

Scaling Mechanics

Since the search API of opensearch and elastic search are compatible, I imagine the mechanics could be very similar to the existing elasticsearch scaler, but rather than elasticsearch client, an open search client could be used instead.

User would need to define:

• url and credentials
• index, search query, parameter
• threshold values

Authentication Source

All supported by the opensearch client: basic, openid token, certificate

Anything else?

No response