[labkey-adam]

Rationale

Wrap all usage metrics supplier dereferences with a helper method that:

• When in dev mode, determines execution time and logs slow pokes (> 1 second)
• Catches and reports exceptions on a per-supplier basis

Introduce NightlyRecalculatingSuppliers, which provides caching suppliers that persist and recalculate nightly. Use them to speed up five slow usage metrics suppliers.

Related Pull Requests

• Use suppliers for every metrics provider #8253

[labkey-jeckels]

This seems fine but also not really that necessary. In typical operations the server calculates metrics once a day in a background thread, so shifting to a maintenance task is OK but not a win in terms of total work.

[labkey-adam]

This seems fine but also not really that necessary. In typical operations the server calculates metrics once a day in a background thread, so shifting to a maintenance task is OK but not a win in terms of total work.

Agreed, for production deployments. This is more about speeding it up for local testing and development (e.g., when adding new metrics). This reduced metrics generation locally from a somewhat annoying ~30 seconds to ~3 seconds.