Pranor Pulse

docker run -p 9090:9090 ghcr.io/vyuvaraj/pranor-pulse:latest

Pranor Pulse is a full-featured, enterprise-grade message broker for the Pranor ecosystem. It supports server-side STOMP brokering, browser-local OPFS-backed queueing, multi-protocol adapters (Kafka wire, MQTT v5), and advanced security (FIPS 140-3, post-quantum cryptography, blind E2EE).

Table of Contents

• Key Features
• Architecture
• API Endpoints
• Protocols Supported
• Browser / OPFS Features
• Security
• Observability
• Kubernetes & Distribution
• Getting Started
• Enterprise Edition

Key Features

📨 Core Broker

• STOMP 1.2 message broker: Topic/queue routing with fan-out, competing consumers, and durable subscriptions
• Exactly-once delivery semantics: Idempotent message IDs with deduplication window
• DLQ + Exponential Backoff Engine: Failed messages automatically moved to Dead Letter Queue with configurable retry policies; exponential backoff with jitter
• Point-in-time event replay: Replay messages from any historical offset on demand
• Schema Registry & Validation: Embedded schema registry for message contract enforcement (Avro/JSON Schema/Protobuf); schema evolution with compatibility checks
• Atomic Multi-Topic Transactions: ACID-style multi-topic publish/consume transactions
• Cooperative Consumer Rebalancing: Sticky partition assignment with graceful rebalance on consumer join/leave

🌐 Browser & OPFS (Local-First)

• OPFS Storage Driver (pkg/opfs): Full browser-native persistent queue using Origin Private File System
• WASM/JS FFI bindings (@pranor/pulse-wasm): Use Pranor Pulse from the browser with a TypeScript SDK
• SharedWorker multi-tab coordination: Single broker across all browser tabs via SharedWorker
• Multi-tab OPFS leader election: navigator.locks-based lease protocol ensures only one tab acts as queue leader at a time
• Client-side AES-256-GCM encryption at rest: Messages encrypted before writing to OPFS
• WebTransport HTTP/3 QUIC relay: Browser outbox relay over QUIC for low-latency connectivity
• Offline outbox & reconnect relay: Queue messages offline; auto-relay when connectivity restores
• Auto-compaction & quota manager: Automatic OPFS quota management with configurable size limits
• Client-side WASM stream filters: Run sandboxed WASM modules to transform/filter messages in-browser
• Persistent storage eviction safeguard: Priority-based eviction prevents silent data loss at storage limits

🗜️ Storage & Compaction

• Write-Ahead Log (WAL) with corruption recovery and CRC checksums
• Topic Log Compaction Policy Engine: Key-based compaction (retain only latest value per key), tombstone purging, TTL-based retention
• Tiered cloud storage offloading: Hot/warm/cold tier management with S3-compatible backend
• Automated storage tiering & compaction: Background compaction scheduler with configurable policies

📡 Protocol Adapters

• Kafka Wire Protocol Compatibility: Drop-in replacement for Kafka consumers/producers (Kafka binary protocol)
• MQTT v5.0 IoT Gateway: Full MQTT v5 adapter — QoS 0/1/2, retain, session persistence, will messages

🔁 Streaming & CDC

• Change Data Capture (CDC) Engine: Database change event streaming (row-level insert/update/delete events)
• Real-time Stream SQL Windowing: Tumbling, sliding, and session windows with aggregations (COUNT, SUM, AVG)

🏢 Multi-Tenant

• Multi-tenant VHosts & rate quotas: Isolated virtual hosts per tenant with per-tenant rate limits and storage quotas
• Zero-Trust OAuth2 & SPIFFE auth: Per-connection authentication with SPIFFE workload identity attestation

Architecture

┌─────────────────────────────────────────────────────────────┐
│                       Pranor Pulse                              │
│                                                             │
│  ┌──────────────┐  ┌──────────────┐  ┌──────────────────┐  │
│  │ STOMP Broker │  │ Kafka Compat │  │  MQTT v5 Gateway │  │
│  └──────┬───────┘  └──────┬───────┘  └────────┬─────────┘  │
│         └─────────────────┼──────────────────-─┘            │
│                           ▼                                 │
│  ┌────────────────────────────────────────────────────────┐ │
│  │              Schema Registry & Validation              │ │
│  └────────────────────────────┬───────────────────────────┘ │
│                               ▼                             │
│  ┌────────────────────────────────────────────────────────┐ │
│  │    WAL Storage Engine │ Compaction │ Tiered Offload     │ │
│  └────────────────────────────────────────────────────────┘ │
│                                                             │
│  ┌─────────────┐  ┌──────────────┐  ┌────────────────────┐ │
│  │ DLQ Engine  │  │ CDC Streamer │  │ SQL Window Engine  │ │
│  └─────────────┘  └──────────────┘  └────────────────────┘ │
└─────────────────────────────────────────────────────────────┘

API Endpoints

Protocols Supported

Browser / OPFS Features

Install the browser SDK:

npm install @pranor/pulse-wasm

import { Pranor Pulse } from '@pranor/pulse-wasm';

const queue = new Pranor Pulse({ encryption: 'aes-256-gcm' });
await queue.publish('orders', { id: 1, item: 'Widget' });
await queue.subscribe('orders', (msg) => console.log(msg));

// Auto-syncs to server when online; stores locally when offline
await queue.enableOfflineSync({ serverUrl: 'wss://queue.pranor.net' });

Security

Observability

• Prometheus /metrics: Per-topic message rate, consumer lag, DLQ depth, compaction stats
• OTel W3C Trace Context: traceparent header propagated per message through full pipeline
• Pranor Console Queue Inspector: Live topic browser, consumer group lag dashboard, DLQ browser with one-click replay, schema registry browser

Kubernetes & Distribution

# Standalone daemon
pranor-pulsed --port 9090 --storage ./data --tls

# CLI
pranor-pulse publish orders '{"id": 1}'
pranor-pulse consume orders --group my-service
pranor pulse publish orders '{"id": 1}'   # Pranor integration

# Kubernetes Operator
kubectl apply -f pranor-pulsecluster.yaml

# KEDA auto-scaling
kubectl apply -f keda-scaledobject.yaml   # Scale consumers on lag

Multi-language client SDKs: Go, TypeScript/JS, Python, Rust, Java.

Cross-cloud active-active geo-replication with automated failover and conflict resolution.

Getting Started

docker run -p 9090:9090 \
  -e PRANOR_PULSE_STORAGE_PATH=/data \
  -e PRANOR_PULSE_OTEL_ENDPOINT=http://pranor-trace:4318 \
  -v queue-data:/data \
  ghcr.io/vyuvaraj/pranor-pulse:latest

Environment Variables

Enterprise Edition