Back to catalog
HyperDX
MITHyperDX replaces proprietary APM platforms with open-source observability, unifying session replays, logs, and traces on ClickHouse and OpenTelemetry.
Open-source alternative to — compare it against:
Overview
HyperDX pairs ClickHouse columnar storage with OpenTelemetry collectors to unify logs, distributed traces, metrics, and session replays under a single interface. Serving as the user interface of ClickStack after ClickHouse's acquisition of the project, it receives telemetry directly through standard network ports without proprietary agent software. By querying structured events in database tables, it provides sub-second diagnostics while keeping data on private infrastructure.
Why HyperDX instead of New Relic
New Relic and Datadog APM rely on proprietary host agents, custom span formats, and metered ingestion fees that penalize high-volume logging. HyperDX replaces that vendor lock-in with an OpenTelemetry-native architecture. Services export traces, metrics, and logs over standard OTLP protocols directly to an ingestion collector, removing proprietary agents from production runtimes. Crucially, HyperDX correlates frontend session recordings with backend traces. When a browser exception occurs, developers inspect the recorded user session alongside the downstream database queries and service spans.
The trade-off involves third-party cloud integrations. New Relic supplies turn-key connectors for managed cloud services and global synthetic probes. HyperDX delegates collection to OpenTelemetry collectors, requiring DevOps engineers to configure scrapers and pipeline routing manually. Storing raw telemetry directly in ClickHouse gives teams schema control, eliminates SaaS egress fees, and enables direct SQL queries without exporting data to external warehouses.
What it does well
- Native OpenTelemetry ingestion: DevOps engineers point standard OTLP collectors directly to ports 4317 or 4318, eliminating vendor runtime agents across Kubernetes clusters.
- Frontend-to-backend trace correlation: Developers link browser session replays to server spans, tracing user actions through downstream microservices in one view.
- Fast log querying: ClickHouse storage enables teams to search billions of structured log lines and trace attributes using property filters without configuring Elasticsearch clusters.
- Live event streaming: Engineers inspect streaming logs and traces in real time during deployments to catch rollout regressions early.
- Automated pattern aggregation: The interface clusters repetitive log lines into distinct groups, helping developers isolate uncommon errors without custom regex.
