Prometheus project rolls out latest version of container management platform
The open source Prometheus project debuted the latest version of its container management platform, version 2.0. The new version includes an independent time series database, improved performance, consistent query latency, staleness handling, built-in support for snapshot backups of databases, and a new storage subsystem.
Prometheus is hosted by the Cloud Native Computing Foundation (CNCF). Similar to Kubernetes, Prometheus is an open source monitoring system for developers. It was designed to keep tabs on containerized environments, leveraging metric tools based on Google’s internal monitoring tools. The technology has been adopted by several companies, including Digital Ocean, Ericsson, CoreOS, Weaveworks, Red Hat, Google, among others.
Among the most significant improvements to Prometheus 2.0 is the independent time series data base, which divides data by ranges of time instead of by data source. This results in query latency that is far more consistent and manageable. According to CNCF, in comparison to Prometheus 1.8, the latest version reduces CPU use by 20 to 40% and disk space usage by 33 to 55%. Disk I/O without much query load is usually less than 1% on average.
An issue Prometheus 2.0 aims to address is how the system handles data. In the past, for example, users were at times flooded with alerts about a server that had gone offline after it had gone back online. The latest version deals with this issues with rules for managing events with resources that have gone stale. “With the new improvements, disappearing monitoring targets or series from those targets are now explicitly tracked, which reduces querying artefacts and increases alerting responsiveness,” wrote Fabian Reinartz in a company blog post.
Prometheus 2.0 also includes support for snapshot backups for the entire database, making it easier to take point-in-time snapshots of the database. In addition, the new storage system is intended to simplify the creation of custom extensions, such as dynamic retention policies. Commenting on the new storage system, Reinartz noted:
“The simple and open storage format and library also allows users to easily build custom extensions like dynamic retention policies. This enables the storage layer to meet a wide array of requirements without drawing complexity into Prometheus itself; allowing it to focus on its core goals.”