Context

Epic SWATCH-4616 describes the problem: when the nightly tally runs, it swamps the queue and disrupts hourly tallies. This causes:

• Extraneous alerting (e.g., paging once every few days)
• Delays in sending usage metrics to metering marketplaces
• Hourly tally messages backing up behind the nightly run

Objective

Investigate and recommend the best approach to create separate processing for hourly tally compared to nightly tally, so that hourly tallies are not disrupted when the nightly run executes.

Deliverables

• Research findings on options (e.g., separate queues, priority handling, scheduling changes)
• Recommended approach with pros/cons
• High-level implementation considerations