Skip to main content
Engineering, Data / ML

Uber: GC Tuning for Improved Presto Reliability

11 January / Global
Featured image for Uber: GC Tuning for Improved Presto Reliability
Image
Figure 1
Image
Figure 2: Presto Full GC count per day.
Image
Figure 3: Heap is divided into regions.
Image
Figure 4: Heap regions are categorized as young gen, old gen, or free.
Image
Figure 5: Heap is fully divided into all the mentioned types.
Image
Figure 6: Full GCs cumulative count of a single cluster
Image
Figure 7: Cumulative Old generation count per day for all clusters.
Image
Figure 8: Internal errors per day on all clusters.
Cristian Velazquez

Cristian Velazquez

Cristian Velazquez is a Staff Software Engineer on the Maps Production Engineering team at Uber. He works on multiple efficiency initiatives across multiple organizations. He has done several tuning across multiple services and multiple Java versions.

Vineeth Karayil Sekharan

Vineeth Karayil Sekharan

Vineeth Karayil Sekharan is a Senior Software Engineer on Uber’s Data Analytics team. He works on managing Presto and is involved in initiatives like JDK version upgrades and Platform Modernization for Presto.

Posted by Cristian Velazquez, Vineeth Karayil Sekharan