Improve metrics loading time about aifororcas-livesystem

[micya]

@micowan - thoughts on above?

[micowan]

You and I are thinking along the same lines. I actually started prototyping something similar to option 3; but ran out time to finish before the beginning of the Hackathon. I had built the metrics collection into the API (vs Azure Functions) as I did not know if Azure Functions would be confusing for the next person to come along and maintain the code. But overall, I like the 3 possible options.

[micya]

You and I are thinking along the same lines. I actually started prototyping something similar to option 3; but ran out time to finish before the beginning of the Hackathon. I had built the metrics collection into the API (vs Azure Functions) as I did not know if Azure Functions would be confusing for the next person to come along and maintain the code. But overall, I like the 3 possible options.

That would work fine too. Will that also count new candidates (i.e., newly flagged but not yet reviewed)?

[micowan]

That was the one I could not crack with the API off the top of my head. Now since we would not have to include the metrics for Tags, Comments, etc. we might not need the aggregate records for unreviewed and could pull those with just a simple change to the index. But I am certainly no Cosmos DB expert, and am open to other solutions. Azure functions may be the only really good way to solve this one.