Mobelium evaluated the yCrash tool for troubleshooting Spark-related OutOfMemoryErrors, transitioning from manual GC log analysis to automated report generation. The tool saved time in root-cause analysis and improved user understanding, especially with recommendations. While beneficial for experienced users, beginners find it challenging. The customer is satisfied and open to exploring additional tools.
A major insurance company improved application performance by tuning Garbage Collection (GC). They identified consecutive Full GC events as the bottleneck and explored three solutions. By increasing the heap size and adjusting GC settings, they significantly enhanced GC efficiency, resulting in a 23% increase in application throughput, a 15% reduction in response time, and a 50% decrease in CPU usage.
The post discusses optimizing a Java application used for controlling warehouse robots, which faced performance issues due to long Garbage Collection (GC) pauses. By analyzing the GC log, it identified a large heap size and the CMS GC algorithm as culprits. Switching to the G1 GC algorithm reduced GC pauses significantly, enhancing application performance without major structural changes.
Sridhar Vembu, CEO of Zoho, inspires many with his success story in building a major SaaS business. He emphasizes the financial benefits of optimizing automatic garbage collection, which can save companies billions annually by reducing application pause times that hinder performance and inflate cloud costs, as shown by successes at Uber and an automobile company.
Uber enhanced compute capacity efficiency by implementing Go GC tuning, saving 70K cores across 30 services. They used a self-referencing finalizer to reduce CPU overhead, achieving significant CPU utilization improvements, notably 65% in observability and 30% in Uber Eats. GC tuning is vital for memory management and optimizing application performance.
Prabhakar Jonnalagadda, a performance architect at Oracle, tackled an application's frequent crashes and poor performance by analyzing its GC log with GCeasy and optimizing JVM arguments. He emphasizes that tuning options are case-specific and encourages engineers to explore and adapt solutions, as there is no universal fix for performance issues.
A leading automobile manufacturer enhanced their middleware platform's response time by 49.46% through optimizing garbage collection with the GCeasy tool, reducing average response time from 1.88 seconds to 0.95 seconds. This adjustment, achieved without code modifications, also decreased transactions exceeding 25 seconds from 0.7% to 0.31%.
This article discusses the optimization of garbage collection (GC) performance to enhance application response time and reduce cloud costs. Key performance indicators include GC pause time, throughput, and CPU consumption. A successful tuning case involved reducing the young generation size from 20GB to 1GB, significantly improving GC throughput and reducing average pause time.
Dsquare is an FX trading boutique focused on short-term market outperformance using algorithmic strategies in forex trading. Jad Sarmo discusses building a low latency Java application that surpassed C++ in performance. He highlights challenges faced and solutions found, including using the GCeasy tool for optimizing garbage collection.