Skip to main content
CometChat In a Box is a production-ready deployment of the CometChat platform that ships with a hardened microservices stack, observability, and scaling patterns. It is designed for teams that need predictable performance, low latency, and operational playbooks for MAU growth from 10k up to 250k+.

​
Who this guide is for

  • DevOps and SRE teams owning uptime and operations
  • Platform, cloud, and backend engineers deploying or tuning the stack
  • Architects evaluating multi-region or compliance-heavy environments

​
Core capabilities

  • Real-time messaging for 1:1 and group chat with durable history
  • WebSocket event streaming for presence, typing, and receipts
  • Kafka-backed event pipeline for decoupled microservices
  • Notifications subsystem for async push delivery and fan-out
  • Moderation services with rule-based filtering and optional AI adapters
  • Webhooks engine with retries and signature validation
  • Horizontally scalable REST APIs for chat, users, groups, and metadata

​
Data & storage summary

  • TiDB cluster (PD, TiKV, TiDB SQL) for primary relational storage
  • MongoDB for flexible metadata and moderation data
  • Redis clusters for caching, pub/sub, sessions, and ephemeral state
  • Kafka as the event backbone
  • Optional object storage (S3, MinIO, Ceph) for media, artifacts, and backups; use when handling large or unstructured objects across services

​
Deployment models

  • Local development (Docker Compose): single-machine setup for development, QA, and CI validation. Not for production use.
  • Docker Swarm (recommended to ~200k MAU): current reference architecture with simple cluster management, secure overlay networking, and rolling updates.
  • Kubernetes (enterprise / >200k MAU): for advanced autoscaling, multi-region failover, service mesh, and regulated environments. Contact CometChat for enterprise architecture guidance.

​
High-level architecture

  • NGINX for TLS termination, routing, WebSocket upgrades, and load balancing
  • WebSocket gateway for presence, sessions, and low-latency delivery
  • Chat API for messaging logic, users, groups, and metadata
  • Moderation engine for rule-based filtering and compliance checks
  • Notifications service for async push workflows
  • Webhooks service for outbound callbacks with retries
  • Kafka for inter-service events and pipelines
  • TiDB, MongoDB, and Redis for stateful data stores
  • Observability stack (Prometheus, Grafana, Loki/ELK) for metrics, dashboards, and logs
  • Private overlay networks to isolate backend traffic