Worker OOMKilled due to insufficient memory limits

Worker pods get OOMKilled before processing tasks. Pods show OOMKilled status with restarts. Memory limits are set too low for worker memory consumption, particularly for memory-intensive tasks.

Check resource usage with kubectl top pods and review current limits with kubectl get deployment. Increase memory limits (recommended minimum 512Mi). For memory-intensive tasks, also configure worker_max_memory_per_child to 200000 (200MB) to restart workers after hitting threshold, and set worker_prefetch_multiplier to 1 to prevent prefetching too many tasks.