Has anyone successfully run Docling on Databricks?
Asked by
MellowOrbit42
My team has started using Docling for document parsing in Databricks. The results are promising, but we occasionally run into GIL-related runtime failures and instability. Has anyone found a reliable way to deploy or scale Docling with Databricks Agent Bricks, or are there better deployment patterns or alternatives we should consider?
Answers
We ended up running Docling through a serving layer rather than directly inside the notebook or job process. That made the setup more scalable, reliable, and cost-effective, especially when several parsing requests were running at once.
Reply to this answer
We ran into similar problems. Processing smaller batches and spreading the workload across more workers helped keep the GIL from locking up the runtime. It still needs some monitoring, but the approach is workable for moderate workloads.
Reply to this answer
The GIL failures seem to be a tradeoff for Docling’s parsing capabilities. If you stay with it, isolate parsing into separate processes or services and keep each batch relatively small rather than running everything in one long-lived Databricks process.
Reply to this answer
You may also want to benchmark alternatives instead of assuming Docling is the best fit. In our testing, Kreuzberg, now called Xberg, was significantly faster and more reliable for our documents, with better task-level results.
Reply to this answer
Add your answer
Answers are reviewed before publication.
Related Discussions