October 2017 Summaries

Oracle's Java 9 Hotspot VM features the Garbage First (G1) Garbage Collector as its default, a system renowned for efficiently handling large heaps with minimal pause times. Introduced in Java 7, the G1 GC works by dividing the heap into numerous small regions, allowing it to collect memory more selectively and concurrently, without completely halting application threads. This contrasts with traditional garbage collectors that classify heaps into young and old generations, requiring more comprehensive sweeps. G1's approach enables it to outperform other collectors in managing large heaps by focusing on regions filled with garbage, thus minimizing workload. However, the G1 GC underperforms when dealing with smaller heaps, where it might trigger "Full GC" cycles if the available heap is insufficient. Despite its innovative design, G1 is not a real-time collector but offers better pause time management than its predecessors. The system's configurability allows it to limit maximum pause times, a feature not feasible with older, more rigid garbage collectors.

The update introduces several enhancements and new features across various platforms and tools, including support for Java 9 and Node.js 8.4, alongside improvements in real user monitoring for JavaScript with MooTools 1.6 compatibility. It also addresses general system improvements, such as setting limits for OneAgent log folders and enhancing cloud infrastructure monitoring. The update includes performance enhancements for the Docker plugin, with support for executing commands without forking in docker-entrypoint.sh and runc, as well as support for Ubuntu node images in Google Container Engine (GKE). Additionally, OneAgent network monitoring sees improvements in the calculation of Response Throughput and Handshake round-trip metrics, and security is boosted by encrypting proxy passwords in the config.properties file.

The webinar "BizDevOps Done Right!" explored the integration of business, development, and operations to enhance innovation cycles and reduce time-to-market for new features, thereby improving end-user experiences. BizDevOps aims to dismantle silos between development, operations, and business, ensuring that end-user feedback is integrated into the development loop to drive business success. Speakers Andi Grabner and Klaus Enzenhofer illustrated the importance of monitoring and feedback loops using real-world examples, emphasizing the need for tools like Dynatrace's AI-powered monitoring to manage technical complexity and improve code quality. The session highlighted the significance of monitoring tools in identifying critical issues and enhancing visibility across the application delivery pipeline, ultimately fostering collaboration between development, operations, and business teams for a seamless and efficient workflow.