Perses is a Cloud Native Computing Foundation sandbox project that functions as a comprehensive dashboard tool for observability visualization. Written primarily in Go, it integrates multiple observability data sources including Prometheus for metrics, Tempo for traces, Loki for logs, and Pyroscope for profiling, consolidating all four observability pillars into a single unified interface. The project is actively maintained with a demo available at demo.perses.dev where users can explore existing dashboards or create their own resources.
The repository demonstrates significant development activity with 708 tracked issues and pull requests. The median response latency for these items is 0.0 hours with a mean of 133.2 hours, indicating responsive community engagement. The most active issue labels are ui with 373 occurrences, priority/P2 with 215 occurrences, and plugin with 153 occurrences, reflecting the project's focus on user interface development and extensibility. The core contributor team includes AntoineThebaud with 1324 recorded events, Nexucis with 854 events, and celian-garcia with 502 events, demonstrating sustained leadership and involvement.
Beyond basic dashboard functionality, Perses pursues several strategic objectives. It aims to establish an open specification for dashboards to foster interoperability across observability tools. The project provides npm packages enabling developers to embed panels and dashboards into their own user interfaces, potentially enhancing visualization capabilities in tools like the Prometheus UI. Extensibility is a core design principle, with support for multiple plugin types that allow users to customize the tool for specific requirements. The project emphasizes GitOps compatibility through SDKs in Golang and Cuelang, a command-line interface for API interaction, static validation tools, and native CLI support to enable Dashboard-as-Code workflows. Additionally, Perses supports Kubernetes-native deployment through Custom Resource Definitions via the separate Perses Operator project.
The project has reached a usable state as an application with a stable data model and multiple panel types covering most monitoring and tracing scenarios. Authentication and authorization features are available. The plugin architecture enables externalization of both plugin loading and implementation, with core plugins maintained in a separate perses/plugins repository. Installation options include precompiled binaries from GitHub releases, Docker images on Docker Hub, Homebrew packages for macOS and Linux users, and building from source using Go 1.23 or greater and NodeJS 22 or greater.
The repository maintains connections with related observability projects including thanos-io/thanos, kiali/kiali, and victoriametrics/victoriametrics through overlapping contributors. The codebase is licensed under Apache 2.0, and each release includes Software Bill of Materials artifacts in SPDX JSON and CycloneDX JSON formats documenting dependencies for Go releases, source trees, and the React frontend. Development contributions are guided by CONTRIBUTING.md documentation, with specific UI development instructions available in the ui/README.md file including architecture details and development guidelines.
