Highlight.io — Open Source Full-Stack Application Monitoring

Introduction

Highlight.io is an open-source observability platform that unifies session replay, error tracking, log management, and distributed tracing under a single interface. It is OpenTelemetry-native and can be self-hosted or used as a managed cloud service. Engineers use it to correlate frontend user sessions with backend errors and traces.

What Highlight.io Does

• Records user sessions with DOM replay to visually reproduce bugs in context
• Captures frontend and backend errors with stack traces and user context attached
• Ingests structured logs via OpenTelemetry or SDK integrations
• Collects distributed traces that link frontend actions to backend service calls
• Provides alerting and dashboards for monitoring error rates, latency, and log patterns

Architecture Overview

Highlight.io uses a Go backend that ingests data via OpenTelemetry (OTLP) and custom SDK protocols. Session replay data is captured by a JavaScript SDK that records DOM mutations and network requests, then batches them to the collector. Errors and logs are routed through the same pipeline. Data is stored in ClickHouse for analytics queries and object storage for replay payloads. The frontend is a React app that renders session replays, error groups, log viewers, and trace flame graphs. Kafka handles event streaming between ingest and storage layers.

Self-Hosting & Configuration

• Deploy with Docker Compose using the provided docker-compose.yml for local or production environments
• Configure the highlight.run SDK in your frontend with your project ID and service name
• Add backend SDKs (Go, Python, Node.js, Ruby, Java) for error and log ingestion
• Point OpenTelemetry exporters to the Highlight collector endpoint for trace ingestion
• Set up alerting rules in the dashboard for error spikes, latency thresholds, and log patterns

Key Features

• Unified view of session replay, errors, logs, and traces in a single timeline
• OpenTelemetry-native ingestion for vendor-neutral observability data
• DOM-based session replay with network, console, and performance waterfall
• Self-hostable with Docker Compose or deployable on Kubernetes
• SDKs for JavaScript, Python, Go, Ruby, Java, and .NET

Comparison with Similar Tools

• Sentry — error tracking with session replay; Highlight adds logs and traces natively
• LogRocket — session replay focused; Highlight includes full backend observability
• Datadog — comprehensive APM but proprietary and expensive; Highlight is open source
• Grafana stack — modular (Loki, Tempo, Grafana); Highlight is a single integrated product
• PostHog — product analytics with session replay; Highlight focuses on engineering observability

FAQ

Q: Is the self-hosted version fully featured? A: Yes. The open-source self-hosted deployment includes all features. The cloud version adds managed hosting, higher retention, and support tiers.

Q: How much storage does session replay use? A: Session replay captures DOM diffs, not video, so storage is compact. A typical session is a few hundred kilobytes of compressed mutation data.

Q: Can I use Highlight with my existing OpenTelemetry setup? A: Yes. Highlight accepts standard OTLP traces and logs, so you can send data from any OpenTelemetry-instrumented service.

Q: Does it affect frontend performance? A: The JavaScript SDK is asynchronous and batches events to minimize overhead. Impact on page load and interaction latency is negligible in production.

Sources

• https://github.com/highlight/highlight
• https://highlight.io/docs