Jaeger Emerges as Meister of Cloud Monitoring
An open-source tool used to monitor and troubleshoot cloud-native services has joined the ranks of maturing platforms such as Kubernetes as microservices scale.
Jaeger, a distributed tracing platform used to track the performance of transactions running on microservices, has “graduated” from incubator to production, the Cloud Native Computing Foundation announced this week. It becomes the seventh project approved by CNCF, joining the de facto standard Kubernetes cluster orchestrator, the Prometheus monitoring system and other cloud-native tools.
With more than 1,000 GitHub pull requests over the last year, Jaeger has emerged as “a battle-tested tool for cloud native monitoring,” said Chris Aniszczyk, the foundation’s CTO/COO.
Jaeger was created by the rideshare service Uber (NYSE: UBER) in 2015 and has since been integrated into thousands of microservices used in production. Along with Uber, Jaeger is being used in production by Northwest Mutual, Weaveworks and Red Hat. It’s also integrated with Red Hat’s OpenShift Service Mesh, CNCF noted.
Jaeger became a CNCF incubator project in September 2017 with 14 releases since then adding features ranging from instrumentation APIs to backend storage.
“Instrument first, ask questions later,” said Goutham Veeramachaneni of Grafana Labs’ Cortex team. Grafana has been using Jaeger on its “observability platform” to track degraded performance. The company and others have contributed improvements, including integration with Prometheus and deployment with Kubernetes clusters.
Jaeger provides a means of determining after the fact what’s happening inside microservices, said Bryan Boreham, engineering director at Weaveworks. “If you don’t have a microscope, you can’t see bacteria. If you don’t have a telescope, you can’t see Jupiter. Jaeger is like that—both the big picture and the fine-grained details at the same time,” Boreham added.
In a deployment setting, Jaeger is promoted as a way to identify the root cause of a performance problem without having to go back to DevOps teams—a capability referred to as observability. “Before Jaeger, we were doing mostly guesswork,” said Veeramachaneni of Grafana Labs.
Along with root case analysis, the tracing platform can be used for distributed transaction monitoring as well as tweaking performance and latency.
To graduate from incubator status, CNCF projects must demonstrate “thriving adoption,” a structured government process and a vibrant community of contributors. Jaeger is currently being used by more than 20 organizations to monitor production workloads, the foundation said.