Tracer-Cloud/opensre
DevOps & Self-hostingSelf-hosted CLI agent that investigates failed CI runs, traces root causes and proposes fixes for review.
What it does
OpenSRE is a CLI agent for CI/CD reliability. When a pipeline run fails, it reads the failing checks and CI logs, pulls recent commits and deploys, correlates logs, metrics and traces, and reasons across connected systems to separate CI-caused failures from code-caused ones. It then returns an evidence-linked answer, can push a fix to the PR branch for review, and can post a summary to Slack, PagerDuty or Telegram. It is aimed at engineers and platform teams who want faster root-cause analysis without sending raw logs to a third party by default.
The agent masks sensitive identifiers before external LLM calls (optional), runs on your own infrastructure, and lets you choose the model provider. A web gateway can be self-hosted as a container or deployed via the provided AWS AMI path, and a managed gateway is available through OpenSRE Cloud. The project is in public alpha, so APIs and integrations may change.
Install and run
Before starting, open a terminal in a code project; a Git repository is recommended to exercise the full workflow.
On macOS or Linux, install from the project installer and start the CLI:
curl -fsSL https://install.opensre.com | bash -s -- -gh
opensre
On Windows, the README uses PowerShell. The installer fetches the latest build from main without sudo; if opensre is not on PATH, follow the printed PATH instructions or open a new terminal Windows install notes. The first launch signs you in and activates the hosted model.
With no subcommand, opensre opens an interactive REPL (a TTY is required). Describe an incident in plain language and use slash commands such as /help, /status, /cost, /integrations list and /integrations verify for session control and integration checks interactive shell commands. For a single non-interactive turn in a script or CI job:
opensre ask "why is checkout-api slow?"
The headless mode supports stdin prompts, JSON output and tool approvals headless CLI guide. Python users can drive the agent in-process from a source checkout via start_embedded_session, as shown in the Python API guide Python API.
Deployment and integrations
OpenSRE can run as a self-hosted container by building the repo Dockerfile, selecting MODE, and setting LLM_PROVIDER plus its API key; set DATABASE_URL only when the gateway needs shared Postgres-backed state. The AWS path builds a gateway image and deploys it with systemd for the Telegram gateway. Full prerequisites are in DEPLOYMENT.md.
The integration surface is broad: LLM providers (Anthropic, OpenAI, Ollama, Gemini, Bedrock and others), observability platforms (Grafana, Datadog, Sentry, CloudWatch and more), infrastructure (Kubernetes, AWS, Azure, ArgoCD, Helm), databases, data platforms, Git hosting and issue trackers. The README lists Linear, Notion, Teams and Confluence as roadmap items rather than available integrations integrations.
Constraints and privacy
Access to the interactive shell requires an active account; you can sign out and exit, but the shell only opens for a signed-in account. The embedded Python API requires a source checkout. Telemetry (first-party analytics and Sentry error reporting) is on by default and can be disabled with export OPENSRE_NO_TELEMETRY=1 telemetry. Local transcript handling is the default and raw logs are not exported in bulk; prompt handling and the security model are described in SECURITY.md. The software is Apache-2.0 licensed and marked public alpha, so expect unstable core workflows license.
HotRepo's take
On the GitHub all-language daily ranking dated 2026-11-19, OpenSRE reached 12th place, adding 25 stars that day, with a total of 12,889 stars and 1,891 forks. That is a strong single-day signal for a project created in January 2026 and still labeled public alpha. Because only one ranking record is available, it supports that day's standing rather than a trend. The combination of a shipped CLI, documented Python embedding, self-hosted deployment and a wide integration table makes this a concrete tool to evaluate for incident triage, with the alpha caveat applying to production use.
Written by AI from the project's README, updated . The project's own docs are the source of truth.
Hacker News
You might also like
- openchoreo/openchoreoAn open-source internal developer platform that turns Kubernetes into self-service abstractions for developers and AI agents.
- dagger/daggerTurns build, test and deploy pipelines into typed code that runs the same on a laptop or in CI.
- openobserve/openobserveA self-hosted observability backend that unifies logs, metrics, traces and frontend monitoring in one binary.
- woodpecker-ci/woodpeckerA self-hosted CI/CD engine for teams that need simple pipelines and extensible plugin support.
- grafana/grafanaOpen-source platform for querying, visualizing and alerting on metrics, logs and traces from multiple data sources.