Added a required 'name' argument to 'collect_metrics'.

[elongl]

No description provided.

[linear]

ELE-3405 Change `collect_metrics` to require a `name` argument

[github-actions]

👋 @elongl
Thank you for raising your pull request.
Please make sure to add tests and document all user-facing changes.
You can do this by editing the docs files in the elementary repository.

[haritamar]

Not sure this is the right place, but can we also save metric_type in data_monitoring_metrics?
(not sure what's the right place for this comment)

[elongl]

So my intuition was to do that as well,
and Ofek convinced me that it shouldn't matter how a metric is calculated for the storage of the metric. Regarding which algorithm to use in the cloud, we can view the type in dbt_tests at the configuration. WDYT?

[haritamar]

Hmm the only case where I think it might matter is when the metric definition is changed - e.g. the name remains the same but the metric type is changed.
But that makes me think that it should actually be added to metric_properties. WDYT?

This achieves two things:

1. The metric will be backfilled if the type changes (same as any other configuration change like timestamp_column)
2. In the cloud I implemented logic that only pulls metrics with the most recent metric properties (to ensure we only rely on metrics collected with the most recent configuration - a bit like what happens in get_anomaly_scores_query).