Sutrace vs SigNoz vs ClickStack — an honest 3-way take

TL;DR. SigNoz is the OSS-first option with an active community and the most mature self-host story. ClickStack (HyperDX-on-ClickHouse, the Hyperdx-rebranded stack from ClickHouse Inc.) is the best fit if you already operate ClickHouse and want a thin observability UI on top. Sutrace is the closest to a Datadog-shaped product with EU-residency-by-default, no per-tag pricing, and four signal surfaces (industrial hardware, software infra, web/APIs, AI agents) in one place. None of these is universally "best." This post is the 1,500-word version of which one fits your team.

The forcing function for this post was HN thread #45294103, where user srcreigh wrote: "I tried both. Signoz is pretty sloppily built…" The thread is worth reading in full because the replies are honest in both directions. SigNoz contributors push back; HyperDX/ClickStack folks pile on; people who tried the managed versions weigh in. It's the most concentrated dose of real-user opinion on this slice of the market.

The three products in one paragraph each

SigNoz. Open-source observability platform built on ClickHouse, with metrics/logs/traces in one UI. Self-host or managed. APM, dashboards, alerts, and an exceptionally active GitHub. Their pricing pages and content marketing are unusually direct about Datadog cost dynamics. The product surface is broad; polish varies by area.

ClickStack. ClickHouse Inc.'s observability stack, built around the HyperDX UI (which they acquired) on top of ClickHouse Cloud. Recent launch of managed pricing — yes, that link is SigNoz's commentary on ClickStack pricing, which is itself worth reading. The product is young but the underlying engine is mature. If you already operate ClickHouse for analytics, this is the path of least resistance.

Sutrace. Next.js 16 + Firebase + Cloudflare delivery, ClickHouse + custom storage on the backend, OTel-native, EU-residency-by-default. Four signal surfaces in one dashboard. Pre-revenue, EU-registered. Built specifically around the cardinality cost-attribution problem. We're the Datadog alternative for teams whose bill broke them.

Feature matrix

When SigNoz wins

• You want self-host and you're willing to operate ClickHouse + a Go service.
• You have an OSS mandate from procurement or government.
• You want a single product that covers metrics, logs, and traces with a community you can file issues against.
• You're comfortable with rough edges in exchange for transparency. The HN thread says it plainly: it's "sloppily built" in some areas, but the alternative is a black box. Some teams trade polish for visibility, and that's a real trade.
• Your data plane is small enough that the managed offering's pricing isn't a constraint.

The SigNoz team responds to issues fast and ships meaningful changes weekly. If you want to influence the roadmap, you can.

When ClickStack wins

• You already run ClickHouse Cloud and your data lives there.
• You want a thin observability UI without operating a separate vendor.
• Logs and traces are your primary need; metrics are secondary.
• You've used HyperDX before and like its query model.
• You want to keep observability data in your own ClickHouse warehouse for analytics joins (e.g., joining application logs to product-analytics tables).

The compute + storage pricing model is closer to AWS than to Datadog — predictable if your workload is steady, surprising if it spikes. SigNoz's analysis of ClickStack pricing is the best public breakdown.

When Sutrace wins

• You need synthetics, industrial signals, or AI agent observability in the same product.
• EU residency by default is non-negotiable.
• You're migrating from Datadog and want a cost-attribution layer that prevents the next surprise bill (the cardinality post explains the architectural piece).
• You want a managed product but don't want to be the largest customer of a 5-person open-source company.
• You operate hardware (PLCs, sensors, MQTT) alongside software services. We're the only one of the three that does this.

We're not better than SigNoz at SigNoz's home turf (OSS metrics/logs/traces). We're not better than ClickStack at ClickStack's home turf (ClickHouse-native query). We win when the workload is heterogeneous and predictability matters more than depth-in-one-domain.

The HN thread, fairly summarized

The thread is the canonical "I tried both" post. The summary, with both sides represented:

On SigNoz. Critics say setup is finicky, the UI has rough edges, and certain features (anomaly detection, advanced metric queries) feel half-finished. Defenders point out that it's open-source, it's improving, and the alternative — paying Datadog — costs orders of magnitude more for marginal polish.

On HyperDX/ClickStack. Praised for query speed and a clean log/trace UX. Criticized for being thin on metrics and for tying you tightly to ClickHouse Cloud's pricing curve.

On managed Datadog alternatives generally. The thread surfaces the recurring complaint: managed alternatives often replicate Datadog's pricing pathologies (per-host or per-GB tiers that don't match how engineers think about cost). The teams who liked the alternatives were the ones who'd done the cardinality audit first.

The migration question

If you're moving off Datadog, the practical question is which of these you can be productive on in week two. We covered the OTel migration in detail in migrating from Datadog to OTel — the week-one checklist. The good news: because all three are OTel-native, the collector config is mostly the same. You change the exporter endpoint and you're done.

The dashboards and alerts are where you spend the time. SigNoz uses its own query language; ClickStack uses HyperDX-style search + ClickHouse SQL; Sutrace uses PromQL + saved views. Pick the one your team will be productive on, not the one with the prettiest demo.

Pricing — order of magnitude

Public pricing changes; treat these as April 2026 snapshots and verify before contract.

• SigNoz Cloud. Per-GB ingest tiers; managed by SigNoz. Cheaper than Datadog at most workload sizes; comparable to ClickStack on logs/traces.
• ClickStack. Compute + storage, billed via ClickHouse Cloud. Predictable for steady workloads; the storage retention multiplier is the line item to watch.
• Sutrace. Flat ingest tiers, no per-tag pricing. Public on the pricing page. EU-residency at no premium.

Where each loses

Be specific about the failure modes.

SigNoz loses when the team can't take on the operational tax of self-hosting and the managed pricing at scale lands close to Datadog. It also loses on niche signal types — synthetics, hardware, AI agents — that aren't on the roadmap.

ClickStack loses when you don't already use ClickHouse, when metrics are your primary need, or when you want a turnkey product that includes alerts and dashboards as first-class citizens. It's a great UI on a great database; it's not yet a complete observability product.

Sutrace loses when you want OSS, when you need 800+ first-party integrations on day one, or when your stack is purely software and the multi-domain story is overhead you don't need. We're not trying to be Datadog. If you need Datadog's breadth, Datadog wins for now.

My honest recommendation

If your team has 5 engineers and a tight budget: try SigNoz Cloud or self-hosted SigNoz. The community will help you when you get stuck.

If your data already lives in ClickHouse: ClickStack is the lowest-friction add.

If your bill broke you, you're EU-headquartered, or you have hardware/AI workloads alongside software: try Sutrace. We'll honestly tell you if we're not the right answer.

If you have unlimited budget and need every integration on day one: stay on Datadog. We mean it.

The wrong move is to pick on vibes. Run a one-week trial against each, look at the cardinality view, look at how the alerting language fits your runbooks, and pick the one your engineers don't dread opening.

Closing

There's no universal winner. The market is finally heterogeneous enough that you can pick on fit instead of on fear. That's a good thing for everyone — including Datadog, who'll have to compete on something other than lock-in.

If you want to see Sutrace against your real workload, pricing is public and the trial accepts your existing OTel collector with one config change. We'll show you the cardinality report on day two and you can decide from there.