Multi-agent, AI-assisted DevOps automation driven by a Slack coordinator, LangGraph pipelines, and Kafka events. One run clones your repo, validates devops_orchestra.yaml, indexes code, runs Sonar analysis, builds, generates and runs tests (optionally grounded in Jira), plans Terraform, and deploys—with human approval where needed.
| Area | What you get |
|---|---|
| Entry | Slack (Socket Mode)—no public GitHub webhook required if events land in Slack |
| Pipeline | LangGraph graphs for GitOps+agents and for infra plan → (approve) → deploy |
| State | Shared DevOpsAgentState; Kafka topics for cross-component observability |
| Config | Per-repo devops_orchestra.yaml (validated schema) |
| Doc | Purpose |
|---|---|
| docs/README.md | Documentation index |
| docs/SETUP.md | Install, Docker, .env, Sonar |
| docs/CONFIGURATION.md | Slack, Jira, channels, tokens |
| docs/ARCHITECTURE_1_TO_2.md | Evolution of the coordinator + graphs |
| docs/FEATURES.md | Features and recent additions |
| docs/FUTURE_SCOPE.md | Roadmap ideas |
| .env.example | Environment variable template |
cp .env.example .env
# Edit .env (Slack, Groq, Jira, Sonar, …)
docker compose up --buildInvite the bot to your command channel; send a pipeline trigger (e.g. run request for a branch). Use a pipeline logs channel or webhook for step-by-step output.
coordinator/ # Slack gateway, pipeline handler, failure suggestions
agents/ # GitOps, build, test, code analysis, infra, deploy, …
langgraph_flows/ # Compiled LangGraph pipelines (full pipeline, infra, deploy)
shared_modules/ # State, Kafka, LLM wrapper, utils
flowchart LR
U[Slack users] --> C[Coordinator]
C --> G1[LangGraph: clone → test]
G1 --> K[Kafka]
C --> A[Approve infra in Slack]
A --> G2[LangGraph: deploy]
G2 --> K
C --> L[Pipeline logs channel]
Add your license file if not present. Prefer issues and PRs that keep state and Kafka contracts backward compatible.