The batch prediction job used to run in minutes.As data volume increased, runtime started doubling unexpectedly.Nothing changed in the model code itself.Now it’s becoming a bottleneck in the pipeline.
Decode Trail Latest Questions
Users can log in without issues, and authentication appears solid.Despite that, some users can access data they clearly shouldn’t.I’m trying to figure out where authorization is breaking down and how to fix it properly.
Single-record updates work fine, but bulk updates expose failures. Errors appear that never show up during normal usage. I want to understand why bulk behavior differs so much.
Our integration uses upsert logic, but duplicates still appear occasionally. External IDs are configured, and the payload looks correct. The duplicates are rare but disruptive. I want to understand why this still happens.
I enabled autoscaling to handle traffic spikes.Instead of improving performance, latency increased.Cold starts seem frequent.This feels counterproductive.
BIASed model issue
