Improve performance of `timeseries_metrics_by_artifact_v0`

[ccerv1]

What is it?

This table is too big to query efficiently. We can't even export it to BigQuery. A few things:

• We should look to better partitions and other performance optimizations in the model itself
• We should probably consider not exporting every metric permutation that exists

[linear[bot]]

OSO-683 Improve performance of `timeseries_metrics_by_artifact_v0`

[ryscheng]

IMO, It's like a P3 to optimize performance on BigQuery.

On Clickhouse, there's indices, order_bys, and other built-in optimizations that will help performance without making any explicit changes to the data. (No need to filter or explicitly partition)

What's the use case you're solving for?

[ryscheng]

I think this is a duplicate of https://github.com/opensource-observer/oso/issues/3182