If you've priced out Datadog or New Relic for a growing app, you already know the pain: the bill scales with your traffic, not your budget. HyperDX and SigNoz both exist to fix that — open-source, self-hostable platforms that unify logs, traces, and metrics in one place, built on OpenTelemetry so you're not locked into a proprietary agent. They solve the same problem in slightly different ways. Here's the real difference, not just the marketing page.
What they have in common
Both HyperDX and SigNoz are MIT-licensed, free to self-host with no data caps on the community edition, and both list Datadog, New Relic, and Grafana as the incumbents they're built to replace. Both use ClickHouse under the hood for fast querying over high-volume telemetry data, and both offer a managed cloud tier if you don't want to run ClickHouse yourself. If you're OpenTelemetry-native already, either one drops in without re-instrumenting your app.
HyperDX — logs, traces, metrics, and session replay in one screen
HyperDX goes a step further than pure observability: it bundles session replay alongside logs, traces, metrics, and error tracking, so you can watch what a user actually did right before an error fired — not just read the stack trace. It was acquired into the ClickHouse family, which adds some reassurance on long-term maintenance. Pricing is usage-based with no per-seat or per-host charge: free for 3 GB/month with 3-day retention and 1 user, then $20/month base for 50 GB/month and 30-day retention, additional data at $0.40/GB. HyperDX claims roughly 10x cheaper than Datadog for equivalent usage.
Good for: teams who want session replay bundled with backend observability, not just server-side telemetry. Limits: smaller plugin/integration ecosystem than the incumbents, and the free tier's 1-user cap means you'll hit the paywall fast on a real team.
SigNoz — the more "pure" OpenTelemetry-native option
SigNoz focuses tightly on the classic three pillars — traces, metrics, and logs — correlated in one OpenTelemetry-native datastore, with no data caps on the free self-hosted Community Edition. It's built to handle real production scale, including high-cardinality tracing up to 1M spans per trace. Cloud pricing starts at $49/month plus usage ($0.30/GB for traces and logs, $0.10 per million metric samples), with Enterprise plans from roughly $4,000/month.
Good for: teams that want a focused, no-caps self-hosted core and are comfortable running ClickHouse themselves. Limits: no session replay or error-tracking layer built in, so you're still reaching for a separate tool if you want that; running ClickHouse at scale is genuine ops work.
Side by side
| HyperDX | SigNoz | |
|---|---|---|
| Core signals | Logs, traces, metrics, errors, session replay | Logs, traces, metrics |
| Self-hosted free tier | Free (MIT), no cap mentioned beyond ops effort | Free (MIT) Community Edition, no data cap |
| Cloud entry price | Free 3GB/mo, then $20/mo base | $49/mo + usage |
| Enterprise | Custom, volume discounts | ~$4,000/mo+ |
| Session replay | Built in | Not included |
| Backing | Acquired by ClickHouse | Independently funded, YC-backed |
Verdict
Pick HyperDX if you want session replay in the same tool as your logs and traces — debugging "what did the user actually click before this broke" without switching to a separate product like LogRocket or FullStory. Pick SigNoz if you want a tighter, more classically OpenTelemetry-native stack and don't need replay, especially if your team is already comfortable running ClickHouse in production. Both are a real, credible escape hatch from a Datadog bill that's grown faster than your infrastructure — the choice comes down to whether session replay earns its place in your stack or not.