Vespper is an open-source AI copilot designed for on-call developers, aiming to streamline incident response by providing real-time, contextual insights and root cause analysis (RCA). It integrates directly into Slack and connects with various observability and incident management tools, offering a self-hostable solution for individual use.

How It Works

Vespper leverages Generative AI to automatically analyze production incidents and alerts. It integrates with tools like Datadog, Coralogix, Opsgenie, and PagerDuty, alongside knowledge sources such as GitHub, Notion, and Confluence, to gather relevant context. The system uses LiteLLM Proxy for unified LLM interaction and ChromaDB for its vector database, enabling conversational querying and automated RCA.

Quick Start & Requirements

• Install: Clone the repository and use docker compose up -d. Alternatively, pull Docker images and use provided scripts.
• Prerequisites: Docker and Docker Compose.
• Configuration: Requires setting OPENAI_API_KEY in config/litellm/.env and Slack tokens/secrets in the root .env file.
• Docs: Getting started, Installation video, Slack app guide.

Highlighted Details

• Automatic RCA from production incidents and alerts.
• Slack integration for conversational interaction.
• Integrations with Datadog, Coralogix, Opsgenie, PagerDuty, GitHub, Notion, Jira, Confluence.
• Self-hostable for data ownership and free use.
• Uses ChromaDB for knowledge base visualization via vector-admin.

Maintenance & Community

• Community support via GitHub Discussions and Slack.
• Project built by Dudu and Topaz.

Licensing & Compatibility

• Licensed under Apache 2.0.
• Suitable for individual use; advanced features are under vespper-ee.

Limitations & Caveats

The project is primarily suited for individual use, with advanced features planned for a commercial offering (vespper-ee). Telemetry is enabled by default but can be disabled via .env.