debug-pipeline-log

To diagnose RAG pipeline performance bottlenecks, you can parse debug logs to calculate wall times for embedding, git prefetching, and chunking phases. This identifies timing anomalies and specific issues like git log latency or GPU backpressure.

Your RAG indexing process may hang during git enrichment due to git log latency or queue starvation. Analyzing pipeline debug logs identifies these specific timing anomalies and bottlenecks within the git prefetching phase.

The best way to identify GPU thermal throttling in RAG pipelines is by evaluating metrics across embedding throughput and concurrency utilization. Parsing debug logs detects this backpressure using predefined diagnostic rules.

Yes, you can use debug logs to optimize embedding throughput and concurrency utilization. The logs contain stage profiling tables that calculate wall times and provide specific tuning recommendations to resolve queue starvation.