March 2020 Summaries

The newly introduced service baseline settings in Dynatrace version 1.189 enable users to tailor alerts according to each service's criticality, allowing for a more flexible and accurate monitoring system. By extending the observation period for performance conditions, false-positive alerts from short-lived spikes can be minimized, ensuring that only critical issues demand immediate attention. This system draws parallels with stock market decisions where timing and observation are crucial, emphasizing the importance of balancing quick reactions to significant events with the need to avoid overreacting to transient issues. The update includes settings that can be adjusted globally or per service, distinguishing between slowdown alerts and error detection to suit the specific needs of different services. Additionally, the default timeout period for low-load events has been reduced to five minutes, preventing unnecessary prolonged alerts and helping maintain efficient monitoring by closing issues sooner. This approach underscores the challenge of distinguishing between critical alerts and false positives, highlighting the value of a longer observation window to reduce alert spam on non-critical services while ensuring timely alerts for critical ones.

Predicting the capacity of Kubernetes nodes or clusters accurately requires high observability, a challenge addressed by Dynatrace through its OneAgent technology. This solution provides instant and self-managed observability out-of-the-box, allowing teams to make informed decisions about capacity without extensive developer involvement. The OneAgent Operator, enhanced by the new Helm chart, simplifies deployment and lifecycle management by automating service discovery and tracing, eliminating the need for manual trace coding. This approach is especially beneficial in high-stakes scenarios, like Super Bowl-scale advertising, where rapid and reliable observability is crucial. Dynatrace's advancements, such as its integration with Google Kubernetes Engine (GKE) and use of the Helm chart, offer streamlined deployment and comprehensive observability without imposing an engineering burden, allowing for efficient management of large-scale Kubernetes environments.

Dynatrace has introduced enhanced MongoDB monitoring capabilities, offering users a comprehensive view of MongoDB performance and its impact on applications. This monitoring feature provides detailed metrics, such as current and available connections, assertions, queue operations, active clients, memory usage, and various database-level statistics like objects, indexes, data size, and storage size. The monitoring is designed for MongoDB 2.6 or higher and requires OneAgent 1.89 or higher, with the user needing the clusterMonitor role for the admin database. Configuration can be done globally, applying to all hosts and automatically activating monitoring for any new MongoDB host detected, or locally, where monitoring is set up for specific hosts.

The release notes for Dynatrace OneAgent version 1.187, published on March 10, 2020, detail new support for Java and Go technologies, including WildFly versions 18-26 and various MySQL and PostgreSQL versions. Future changes in operating system support are highlighted, with several Linux and Windows versions facing end-of-support dates ranging from October 2025 to May 2026. The release includes a series of bug fixes across multiple components such as .NET, Java, PHP, and Node.js, addressing issues like CPU overhead during Apache Camel detection and crashes in PHP modules. Additionally, the update resolves problems related to SQL sensor settings in Go, memory reporting on Docker, and DB2 string handling in mainframe environments. The release also introduces early adopter features like continuous memory profiling for Java and enhanced visibility into SQL statements for Go applications, though their general availability dates remain unconfirmed.