The dashboard sync job keeps pulling duplicate wrapUpCode values when I route historical pulls through GET /api/v2/conversations?query=type:voice&intervalStart=2024-01-01T00:00:00Z instead of the analytics bucket. Node 18.17 throws a memory spike. Switching to GET /api/v2/analytics/conversations/summary fixes the bloat but the intervalCount metric stays zero until the aggregator catches up. Can’t figure out the exact rule on when to use the real-time feed versus the analytics endpoint for reporting so I’m just patching the json payloads together manually.