Content by bhaktirath95 (3)

BhaktiRath95 walks through common failure modes when running AI/ML inference workloads on Azure Container Apps, including slow model startup, probe timeouts, OOM kills, and GPU initialization problems. The post provides concrete probe settings, Python/FastAPI patterns, and Log Analytics queries to diagnose and fix issues methodically.

BhaktiRath95 breaks down why Azure Container Apps can feel “slow to start” in production, separating true cold starts from scaling delays and resource throttling. It includes concrete fixes like minReplicas tuning, KEDA rule adjustments, probe configuration, image-size reduction, and practical .NET and Django startup optimizations backed by Log Analytics and Application Insights queries.

BhaktiRath95 walks through common startup and deployment failures in Azure Container Apps and Container App Jobs for .NET and Django workloads, showing what the errors look like in logs, why they happen, and the concrete CLI, configuration, and code changes that fix them.