HyperDX

What is HyperDX?

HyperDX is an open source observability platform that unifies logs, metrics, traces, session replays, and errors in a single search interface. Built on ClickHouse for storage and OpenTelemetry for instrumentation, it lets you follow one production issue from a user’s browser click all the way down to the slow database query behind it. It is MIT-licensed and self-hostable, and was acquired by ClickHouse in 2025.

What is HyperDX best for?

HyperDX is best for engineering teams that want full-stack observability without Datadog- or New Relic-scale bills. It fits teams already comfortable running infrastructure who want to own their telemetry data, and it is especially strong when you want session replays correlated with backend traces — seeing what a user did and what the server did for the same request, in one place.

What can HyperDX do?

• Unified search across logs, metrics, traces, session replays, and errors, correlated by request so you jump between them without switching tools.
• Session replay tied to backend telemetry — watch the user’s actual browser session next to the traces and logs it generated.
• End-to-end tracing from HTTP request through to database queries, for APM-style latency debugging.
• Full-text and property search with a simple syntax (e.g. level:err) — no SQL required, though native JSON and ClickHouse SQL are available.
• Live tail for real-time streaming of logs and traces as they arrive.
• Automatic event/log pattern clustering and high-cardinality dashboards to surface anomalies and trends.
• Alerting to Slack, email, and PagerDuty on any search or dashboard.
• Schema-agnostic querying that works on existing ClickHouse tables without remodeling your data.

Where does HyperDX fall short?

HyperDX depends on ClickHouse, so self-hosting means running and operating a ClickHouse instance (or pointing at ClickHouse Cloud) — that is a heavier, more specialized backing store than a single Postgres, and tuning it for large ingest volumes is its own skill. The project is also younger and narrower than Datadog or Splunk: it has fewer turnkey integrations, prebuilt dashboards, and enterprise features (advanced RBAC, long-retention compliance tooling) than the incumbents it replaces. Finally, some conveniences such as SAML SSO are reserved for the paid cloud’s Enterprise plan rather than the open source build.

Is HyperDX free?

Yes — the self-hosted edition is free and open source under the MIT license, with a Docker all-in-one image to get started; you only pay for the servers you run it on. HyperDX Cloud is the managed option: a free tier stores up to 3 GB/month with 3-day retention for one user, and the Starter plan is $20/month including 50 GB/month at 30-day retention with unlimited users, then $0.40 per additional GB. Pricing is usage-based with no per-seat or per-host fees.

What does HyperDX replace?

HyperDX is a practical open source alternative to hosted observability suites — most directly Datadog and New Relic for full-stack monitoring and APM, Splunk for log search and analytics, and LogRocket for session replay. Because it consolidates all of those into one platform, it can stand in for a stack that would otherwise mean several separate paid tools. Under the hood it stores data in ClickHouse and ingests via OpenTelemetry, so it also overlaps with tracing backends like Jaeger while adding logs, metrics, and replay on top.

FAQ

Is HyperDX open source? Yes. HyperDX is released under the permissive MIT license, and the full platform — including session replay and tracing — can be self-hosted from its GitHub repository at no cost.

Can I self-host HyperDX for free? Yes. The self-hosted edition is free; a Docker all-in-one image is provided, and the docs recommend at least 4 GB RAM and 2 CPU cores for testing. You supply the ClickHouse backend (self-run or ClickHouse Cloud).

Is HyperDX a good Datadog alternative? For teams wanting to own their data and avoid usage-based, per-host billing, yes — it covers logs, metrics, traces, and session replay in one UI. Datadog still leads on breadth of integrations and mature enterprise features, so the trade is cost and data ownership against turnkey coverage.

What do I need to run HyperDX? A host running Docker (or Kubernetes) and a ClickHouse instance for storage, plus OpenTelemetry instrumentation in your apps to send telemetry. Official SDKs cover browser, Node.js, Python and more, and any OpenTelemetry library works.