90597e47ca
- WorkflowCronTriggerCronJob was querying every active workspace (~700) sequentially every minute to find cron triggers, causing regular CPU spikes to 100% on worker pods - Added a Redis hashset cache that stores cron triggers. On cache miss (TTL expired, cold start, or explicit invalidation), a full scan rebuilds the cache - Creating/deleting a new cron trigger updates the cache, only if exists