December 2013 Summaries

Java applications like Tomcat or Cassandra often rely on JMX for performance monitoring. However, extracting metrics from JMX can be challenging and time-consuming. Previously, system admins and developers used JConsole to view these metrics quickly but it had limitations such as not retaining historical data, limited graphing capabilities, and lack of alerting support. Datadog's newly released agent addresses these issues with an upgraded JMX connector called JMXFetch. This tool collects JMX metrics automatically and makes them available for analysis, alerting, and correlation within the Datadog platform. With Datadog, users can graph JMX metrics from specific hosts or aggregated metrics from tagged hosts, set alerts on individual or aggregated metrics, and correlate JMX metrics with events in other systems to understand their impact on application performance.

The latest version of Datadog's agent (dd-agent 4.0.0) has been released, offering new integrations for PostFix and Couchbase. This release focuses on collecting more metrics from existing integrations to provide users with enhanced visibility into their application stack. A new JMX collector is included in the updated version of the Agent, making it easy to customize metrics from various Datadog integrations. The new Agent also collects additional metrics for NGINX, Postgres, Cassandra, Tomcat, Solr, ActiveMQ, and HAProxy. Additionally, Couchbase integration has been added thanks to contributions from @jslatts.

The new Datadog Agent (dd-agent 4.0.0) has been released, providing enhanced visibility into application stacks through its JMX collector and improved metrics collection from existing integrations such as PostFix, Couchbase, NGINX, Postgres, Cassandra, Tomcat, Solr, ActiveMQ, and HAProxy. The new Agent includes features like long-lived connections for NGINX, deadlocks detection in Postgres, temporary file allocation metrics in Postgres, per-column-family basis metrics in Cassandra, generic JMX metric collection methods for other integrations, aggregated backend metrics for HAProxy, and per-bucket metrics for Couchbase. These updates were made possible through contributions from the community and beta testing support.

Last month at AWS re:Invent in Las Vegas, attendees experienced an insightful week meeting innovative AWS users and engaging with the Amazon team. The event included a presentation by Mike Fiedler, Director of Technical Operations, titled "Instrumenting Application Stacks in a Dynamically Scaling Environment." In this talk, Mike discussed the importance of gaining visibility into application stack performance to understand its functioning and the challenges of instrumenting components in a dynamically scaling environment. He then demonstrated how to eliminate manual steps in collecting data from an elastically scaling cloud stack by auto-instrumenting hosts for graphing and alerting purposes. To access the metrics graphs shown during the presentation, sign up for a free trial of Datadog. The event was well-organized by Amazon Web Services.

AWS re:Invent is an annual conference where AWS users meet and learn about the latest developments in cloud computing. Amazon's Director of Technical Operations, Mike Fiedler, delivered a presentation titled "Instrumenting Application Stacks in a Dynamically Scaling Environment" at the event. The presentation focused on the importance of gaining visibility into application stack performance and how instrumenting components can be challenging in dynamically scaling environments. Mike demonstrated a technical approach to remove manual steps by auto-instrumenting hosts for graphing and alerting, showcasing metrics graphs to illustrate different system behaviors. The presentation highlights the need for monitoring tools like Datadog to gain insights into application performance. AWS provided a great experience for attendees at the conference.