AI – Performance Methodology in the Cloud – Part 2 – Harshad

• Flow or FloGo, an open-source, microservice-based blockchain application, had problems with its performance when subjected to a higher load.
• To figure out the root cause of the problem, the developers used performance profiling to identify which parts of the application were causing the bottleneck.
• By tracing the traffic between different nodes, they were able to identify that the consensus nodes were causing the bottleneck.
• Performance characterization includes mapping to understand the part of the application that needs to be investigated, and generating a flame graph to show where the application is spending time.
• The access node is the most non-performant part of the application and has a high CPI.
• To improve performance, the access node should be investigated and optimized.

Learn about how developers used performance profiling to identify and address performance issues in the open-source, microservice-based blockchain application Flow or FloGo. Topics include identifying bottleneck causes, tracing traffic between nodes, performance characterization, and improving access node efficiency.