FAQ

What is OpenSRE?

OpenSRE is an open-source framework for AI SRE agents that investigate production incidents. It uses your observability stack, cloud context, and runbooks; runs a structured root-cause analysis pipeline; and delivers findings to Slack, local files, or other messaging channels. See the Integrations overview for supported tools.

How do I get started?

Install OpenSRE, complete onboarding, then investigate a sample alert:

curl -fsSL https://install.opensre.com | bash
opensre onboard
opensre investigate -i tests/e2e/kubernetes/fixtures/datadog_k8s_alert.json

Full walkthrough: Quickstart.

Which deployment path should I use?

Most users start with the local CLI (curl or Homebrew install, then opensre onboard). For organization-wide hosted connectors, use Enterprise setup at app.tracer.cloud.Self-hosted gateway deployment is covered in Deployment.

Can I self-host OpenSRE?

Yes. Deploy the gateway as a FastAPI application using the repository Dockerfile or your host’s Python workflow. Set LLM_PROVIDER and the matching provider API key (for example ANTHROPIC_API_KEY when LLM_PROVIDER=anthropic). See Deployment and Environment variables.

Which model providers are supported?

OpenSRE supports Anthropic, OpenAI, OpenRouter, TrustedRouter, DeepSeek, Gemini, Bedrock, Azure OpenAI, and CLI-backed providers. Set LLM_PROVIDER and the matching API key. Provider matrix, defaults, and troubleshooting: LLM providers.

Does OpenSRE work with our existing tools?

In most cases, yes. OpenSRE integrates with observability platforms (Datadog, Grafana, Sentry), cloud environments (AWS, Kubernetes), incident tools (PagerDuty, OpsGenie), databases, and messaging (Slack, Telegram). Browse the Integrations overview and run opensre integrations verify after setup.

What happens when I run opensre with no arguments?

Running opensre starts the interactive shell. You can describe issues in plain language, stream investigations, and ask follow-up questions in the same session. Slash commands such as /help, /investigate, and /verify datadog are documented in Shell commands.

How are security and telemetry handled?

OpenSRE supports reversible masking before external LLM calls and command-history redaction via Interactive shell privacy. Anonymous product telemetry can be disabled with OPENSRE_NO_TELEMETRY=1. Report vulnerabilities to support@opensre.com.

Deploy and APIs

Community