September 2018 Summaries

Monitoring Java applications involves analyzing various metrics within the Java Runtime Environment (JRE) to ensure optimal performance and troubleshoot issues. Key aspects include understanding memory usage by profiling stack and heap memory, which are crucial for preventing memory leaks and unexpected service degradation. The stack handles function calls and local variables, while the heap is used for object allocation, both of which need careful monitoring to avoid issues like OutOfMemoryError. Garbage collection processes, such as the Serial, Parallel, CMS, and G1 collectors, play a vital role in memory management by freeing up unused memory, though they introduce overhead and need configuration based on application requirements. Thread monitoring is essential, especially in a concurrent environment, to manage system resources effectively and prevent performance bottlenecks. The use of tools like Sysdig allows for effective monitoring without direct instrumentation of Java applications, offering insights into kernel metadata and events, which is particularly useful in containerized environments like Docker and Kubernetes. The article highlights the importance of understanding JRE metrics and suggests that further exploration into Java Management Extensions and Managed Beans can enhance observability at the application layer.

Sysdig Monitor 3.0 introduces enterprise-grade enhancements for monitoring Kubernetes and Docker environments, emphasizing the integration of Prometheus capabilities to meet the needs of large-scale, complex production settings. The update includes features like Prometheus Query Language (PromQL) compatibility, a Grafana plugin, and enhanced support for Kubernetes, offering universal support across various Kubernetes distributions and new cluster management dashboards. These advancements aim to provide users with scalable monitoring, long-term data retention, and improved visibility and control across multi-cluster and multi-cloud environments. Sysdig Monitor 3.0 also introduces streamlined event feeds and dashboard templates to enhance usability, allowing enterprises to efficiently monitor and manage their cloud-native applications and infrastructure, ultimately accelerating their operational efficiency and time-to-value.

Kubernetes 1.12 introduces a variety of enhancements across multiple domains, including scalability, security, storage, and cloud provider support. Key updates include the beta release of arbitrary custom metrics for the Horizontal Pod Autoscaler, allowing more precise scaling based on specific metrics, and the introduction of Pod Vertical Scaling for dynamic resource allocation. In security, kubelet TLS bootstrap reaches stable status, and certificate rotation for kubelets is now in beta. Storage improvements feature the alpha release of snapshot and restore volume support, while topology-aware dynamic provisioning is now in beta. The update also brings support for Azure Availability Zones and Virtual Machine Scale Sets, improving cloud provider integration. Internally, Kubernetes enhances multi-platform compatibility and introduces an easier configuration process through ComponentConfig. New features like SCTP support for services and a more efficient scheduler improve performance and functionality, making Kubernetes 1.12 a comprehensive update that addresses both user-facing and internal enhancements.

Jorge Salamero Sanz's blog post explores the various methods of code instrumentation, focusing on custom metrics, Application Performance Management (APM), and OpenTracing. The text elaborates on the importance of monitoring applications to prevent issues and enhance user experience, emphasizing that monitoring aids in troubleshooting and maintaining application performance. It contrasts the strengths and limitations of custom metrics and APM, noting that custom metrics provide full stack visibility and are lightweight, while APM offers transaction-level observability but can be resource-intensive. OpenTracing is presented as a standard for distributed tracing, complementing both methods. The choice between these tools depends on specific needs, such as whether the focus is on code-related issues or infrastructure performance. The article also provides insight into implementing custom metrics in various programming languages, highlighting the necessity for developers to understand their application’s behavior to identify and resolve issues effectively.

Sysdig's Falco Feeds enhances the capabilities of Falco, an open-source security tool, by providing companies with access to expertly crafted rules that are regularly updated to address emerging threats. This initiative is designed to support open-source-focused companies in bolstering their security measures. The documentation, published on September 11, 2018, highlights the importance of continuous updates in maintaining effective threat detection and offers resources for organizations to test and improve their cloud defense strategies through expert guidance and demonstrations.