KubeRCA Backend

Go REST API Server for Incident Management & Alert Processing

---

Overview

The KubeRCA Backend is a Go-based REST API server that serves as the central hub for the KubeRCA system. It receives Alertmanager webhooks, manages incidents and alerts, handles authentication, and coordinates with the Agent service for AI-powered analysis.

Key Responsibilities

• Receive and process Alertmanager webhook alerts
• Create and manage incidents with alert associations
• Send Slack notifications with thread tracking
• Coordinate analysis requests with the Agent service
• Store and search incident embeddings via pgvector
• Provide JWT-based authentication

---

Architecture

flowchart LR
  AM[Alertmanager] -->|Webhook| BE[Backend]
  BE -->|Notification| SL[Slack]
  BE -->|Analyze Request| AG[Agent]
  BE -->|Embeddings| LLM[Gemini API]
  BE <-->|Data| PG[(PostgreSQL + pgvector)]
  FE[Frontend] -->|REST API| BE

Loading

---

Tech Stack

Category	Technology
Language	Go 1.24
Framework	Gin
Database	PostgreSQL + pgvector
Auth	JWT (golang-jwt/jwt/v5)
AI	Google GenAI (embeddings)
Container	Docker (multi-stage build)
CI/CD	GitHub Actions

---

Quick Start

Prerequisites

• Go 1.24+
• PostgreSQL with pgvector extension
• (Optional) Slack Bot Token for notifications

Installation

# Run in repository root
# (monorepo layout: cd backend/main)
go mod tidy

Run Development Server

go run .

The server starts at http://localhost:8080 by default.

Run Tests

go test ./...

Build Binary

go build -o main .

---

Merge Policy (release-please)

• Prefer Squash and merge (or Rebase and merge) into main to keep a single conventional commit per change.
• Avoid merge commits with conventional prefixes (e.g., feat:, fix:) to prevent duplicate changelog entries.
• Do not edit CHANGELOG.md manually; release-please owns it.

---

API Endpoints

Health & Status

Method	Endpoint	Description
GET	/	Server status
GET	/ping	Health check (returns "pong")
GET	/openapi.json	OpenAPI specification

Authentication (/api/v1/auth)

Method	Endpoint	Description
POST	/register	User registration (when enabled)
POST	/login	User login
POST	/refresh	Refresh access token
POST	/logout	Logout (revoke refresh token)
GET	/config	Auth configuration
GET	/me	Current user info

Webhook

Method	Endpoint	Description
POST	/webhook/alertmanager	Receive Alertmanager alerts

Incidents (/api/v1/incidents)

Method	Endpoint	Description
GET	/	List incidents
GET	/:id	Get incident details
PUT	/:id	Update incident
PATCH	/:id	Hide incident
GET	/hidden	List hidden incidents
PATCH	/:id/unhide	Unhide incident
POST	/:id/resolve	Resolve incident & trigger final analysis
GET	/:id/alerts	List alerts for incident
POST	/mock	Create mock incident (testing)

Alerts (/api/v1/alerts)

Method	Endpoint	Description
GET	/	List alerts
GET	/:id	Get alert details
PUT	/:id/incident	Reassign alert to different incident
POST	/:id/analyze	Trigger alert analysis
POST	/:id/resolve	Manually resolve alert (Slack + Agent analysis)
POST	/bulk-resolve	Bulk resolve alerts (up to 50, Slack only)

Feedback (/api/v1/incidents/:id & /api/v1/alerts/:id)

Method	Endpoint	Description
GET	/incidents/:id/feedback	Get incident feedback (votes + comments)
POST	/incidents/:id/comments	Add incident comment
PUT	/incidents/:id/comments/:commentId	Update incident comment
DELETE	/incidents/:id/comments/:commentId	Delete incident comment
POST	/incidents/:id/vote	Vote on incident analysis
GET	/alerts/:id/feedback	Get alert feedback (votes + comments)
POST	/alerts/:id/comments	Add alert comment
PUT	/alerts/:id/comments/:commentId	Update alert comment
DELETE	/alerts/:id/comments/:commentId	Delete alert comment
POST	/alerts/:id/vote	Vote on alert analysis

Chat (/api/v1/chat)

Method	Endpoint	Description
POST	/	Send chat message (context-aware AI chat)

Webhook Settings (/api/v1/settings/webhooks)

Method	Endpoint	Description
GET	/	List webhook configurations
POST	/	Create webhook configuration
GET	/:id	Get webhook configuration
PUT	/:id	Update webhook configuration
DELETE	/:id	Delete webhook configuration

Realtime Events (/api/v1/events)

Method	Endpoint	Description
GET	/	SSE event stream (incident/alert updates)

Analytics (/api/v1/analytics)

Method	Endpoint	Description
GET	/incidents	Incident analytics metrics
GET	/summary	Overall summary statistics

App Settings (/api/v1/app-settings)

Method	Endpoint	Description
GET	/	Get app settings
PUT	/	Update app settings

Embeddings (/api/v1/embeddings)

Method	Endpoint	Description
POST	/	Create embedding
POST	/search	Search similar incidents

---

Configuration

Environment Variables

Variable	Description	Required
DATABASE_URL	PostgreSQL connection string	Yes
SLACK_BOT_TOKEN	Slack Bot OAuth token	No
SLACK_CHANNEL_ID	Slack channel for notifications	No
AGENT_URL	Agent service base URL	No (default: http://kube-rca-agent.kube-rca.svc:8000)
AI_API_KEY	Gemini API key for embeddings	Yes
JWT_SECRET	JWT signing secret	Yes
JWT_ACCESS_TTL	Access token TTL (e.g., 15m)	No
JWT_REFRESH_TTL	Refresh token TTL (e.g., 168h)	No
ALLOW_SIGNUP	Enable user registration (true/false)	No
ADMIN_USERNAME	Initial admin username	No
ADMIN_PASSWORD	Initial admin password	No
CORS_ALLOWED_ORIGINS	Allowed CORS origins (comma-separated)	No

Cookie Configuration

Variable	Description	Default
AUTH_COOKIE_SECURE	Secure flag	true
AUTH_COOKIE_SAMESITE	SameSite policy	Lax
AUTH_COOKIE_DOMAIN	Cookie domain	-
AUTH_COOKIE_PATH	Cookie path	/

Local Development

Create a .env file in the backend directory:

DATABASE_URL=postgres://user:pass@localhost:5432/kubereca?sslmode=disable
SLACK_BOT_TOKEN=xoxb-your-token
SLACK_CHANNEL_ID=C01234567
AGENT_URL=http://localhost:8000
AI_API_KEY=your-gemini-api-key
JWT_SECRET=your-secret-key
ALLOW_SIGNUP=true
ADMIN_USERNAME=admin
ADMIN_PASSWORD=admin

---

Database Schema

PostgreSQL with pgvector extension is required. Key tables:

-- Enable pgvector
CREATE EXTENSION IF NOT EXISTS vector;

-- Embeddings table for similarity search
CREATE TABLE embeddings (
  id BIGSERIAL PRIMARY KEY,
  incident_id TEXT NOT NULL,
  incident_summary TEXT NOT NULL,
  embedding VECTOR(3072) NOT NULL,
  model TEXT NOT NULL,
  created_at TIMESTAMPTZ NOT NULL DEFAULT NOW()
);

---

Project Structure

backend/
├── main.go              # Application entrypoint
├── openapi.go           # OpenAPI annotations
├── docs/                # Generated Swagger docs
│   ├── docs.go
│   ├── swagger.json
│   └── swagger.yaml
├── internal/
│   ├── client/          # External service clients (Slack, Agent)
│   ├── config/          # Configuration loading
│   ├── db/              # Database connections & queries
│   ├── handler/         # HTTP handlers (routing)
│   ├── model/           # Data models/DTOs
│   └── service/         # Business logic
├── Dockerfile           # Multi-stage Docker build
└── .github/workflows/
    └── ci.yaml          # CI/CD pipeline

---

Development

Generate OpenAPI Spec

go run github.com/swaggo/swag/cmd/swag@v1.16.6 init -g openapi.go --parseInternal --output docs

Code Formatting

go fmt ./...

Linting & Testing

go fmt ./... && go test ./...

---

Docker

Build Image

docker build -t kube-rca-backend .

Run Container

docker run -d -p 8080:8080 \
  -e DATABASE_URL=postgres://... \
  -e JWT_SECRET=your-secret \
  kube-rca-backend

---

API Testing

Health Check

curl http://localhost:8080/ping
# Response: {"message": "pong"}

Server Status

curl http://localhost:8080/
# Response: {"status": "ok", "message": "Gin basic API server is running"}

---

Related Components

• KubeRCA Agent - Python analysis service
• KubeRCA Frontend - React web dashboard
• Helm Charts - Kubernetes deployment

---

License

This project is part of KubeRCA, licensed under the MIT License. See the LICENSE file for details.