Sutrace vs SigNoz

TL;DR. SigNoz is the leading open-source OpenTelemetry-native observability stack, with both a self-host distribution and a SigNoz Cloud managed offering. Sutrace is managed-only with EU residency by default and a different opinion on what should ship in the box. The two products share an underlying architectural choice (ClickHouse + OTel Collector) and disagree on everything else — license, hosting model, default UX, residency posture. If self-host or OSS is a hard requirement, SigNoz is the right answer and we'll say so on a sales call. If you want managed-with-EU-default and you don't need OSS, this comparison is the substance of the decision.

Side-by-side

The honest framing of SigNoz's self-host story

The most cited critique of SigNoz's self-host distribution is from HN user srcreigh (in the HyperDX vs SigNoz thread):

Signoz is pretty sloppily built. For ex the self hosted option starts a ZK instance with 1 clickhouse host — no way to disable, 800MB ram.

We've quoted this before because it's the canonical practitioner concern. The SigNoz team has been engaged on the GitHub issues that follow from this — the operational footprint is heavier than a small team typically wants for a side-project trial. For production teams that already run ZooKeeper or ClickHouse in their stack, the footprint is fine. For a 5-engineer startup wanting to "spin up observability for a weekend," it's overkill.

This is not a damning critique of SigNoz. It's a fair observation about the cost-of-entry for OSS self-host. The architectural choice (ClickHouse + replication via ZK/Keeper) is the same choice we make at Sutrace; we just hide the operational cost behind the managed service.

If you have the team to operate the footprint, SigNoz self-host is excellent. If you don't, SigNoz Cloud or Sutrace are the managed alternatives.

When SigNoz wins

We're being honest about this set; these are the cases where we'd recommend SigNoz over us on a sales call.

1. OSS is a hard requirement. Compliance, philosophy, contractual — whatever the reason, if your team requires open-source, we are not the answer. SigNoz is the strongest OSS option in the OTel-native category.

2. You need full self-host with no managed service in the loop. Air-gapped, on-prem, regulated industries that can't have data leave their physical infrastructure. SigNoz self-host is the answer; Sutrace is managed-only.

3. You have ClickHouse expertise on the team. SigNoz exposes more of the ClickHouse layer than we do. If your team is comfortable writing schema migrations and tuning ClickHouse settings, you'll extract more value.

4. You're community-first. SigNoz has an active OSS community, public roadmap, GitHub-driven development. Some teams value the transparency that OSS development affords. We ship privately.

5. You're cost-floor sensitive at the very smallest scale. SigNoz self-host on a single machine is genuinely free. Our smallest plan has a price floor. For a side project, free wins.

6. You're philosophically aligned with the OSS-first observability movement. The OTel protocol-war pillar categorises SigNoz alongside other OSS-first OTel-native products (Uptrace, ClickStack, HyperDX). If your reading of the industry is "OSS-first observability is the future," SigNoz is the bet.

When Sutrace wins

1. EU residency is mandatory and you can't operate self-host. SigNoz Cloud's primary region today is US; EU is on roadmap. Sutrace is EU-default.

2. You don't have a platform team to operate ClickHouse + ZK. Managed, no operational footprint, on-call within minutes of provisioning.

3. Cardinality cost attribution as a product, not a query. SigNoz lets you query cardinality. Sutrace makes it the metric's own page, with budget enforcement at the collector. We covered the architecture in cardinality cost attribution before the bill arrives.

4. Tuned-by-default alerting. SigNoz ships minimal default rules; you bring your own. Sutrace ships the 5-rule starter on by default.

5. Mixed signal types beyond software. Industrial (PLC/SCADA via OPC-UA), AI agents (LLM cost/latency), web/APIs (synthetics) — one dashboard, one query layer, one billing axis. SigNoz focuses on traces/metrics/logs for software. The OTel use-case page covers the architecture.

6. You want a vendor relationship. Sales engineers, dedicated CSM at scale, contractual SLAs. SigNoz has these on Cloud; the SLA tier and EU-region story is currently smaller than ours.

Pricing comparison

SigNoz self-host is free in license; you pay for the infrastructure (a ClickHouse cluster, ZooKeeper / Keeper, and the SigNoz services themselves). At small scale this is one VM and ~$50/mo; at production scale this is real engineering effort and can run several thousand dollars per month in raw infrastructure plus the SRE time to operate it.

SigNoz Cloud and Sutrace are both per-GB-ingest with similar pricing structures; the differences are in the bundled features (Sutrace bundles synthetics and a richer alerting feature set; SigNoz Cloud bundles different things). Run both pricing calculators with your actual ingest numbers.

The honest move at scale: if you have the team, self-host SigNoz is structurally cheaper than any managed offering. If you don't, the managed offerings are similar enough that the decision should be on residency, support, and feature fit — not raw price.

Migration paths

SigNoz → Sutrace. Both are OTel-native with ClickHouse-backed storage. The migration is a re-pointing of the OTel Collector's otlp exporter. Dashboards are recreated (different UI). Alerts are recreated. ClickHouse data does not migrate (different schema in places, different tenancy model). Plan for two weeks of focused work.

Sutrace → SigNoz. Same shape, reverse direction. We don't make this hard — your data is in OTel format on the wire and we don't re-tag at ingest. If you decide to leave us, you can. We'd rather you don't, but we're not going to lock you in.

This bidirectional symmetry is one of the upsides of building on shared open standards. It's also one of the strongest arguments for staying off proprietary protocols (Datadog) — there, the migration cost is much higher.

What we won't do

• We will not pretend SigNoz is "sloppy" beyond the specific operational-footprint critique. The SigNoz team is good and the product is solid; the critique is fair on its narrow scope.
• We will not match SigNoz's OSS posture. We don't plan to be open source.
• We will not undercut SigNoz Cloud on every line item. Some workloads it's cheaper.

What to do next

1. Spin up SigNoz Docker Compose for an afternoon. Honestly assess the operational footprint against your team's appetite.
2. If self-host doesn't fit, evaluate SigNoz Cloud and Sutrace in parallel for two weeks. Same OTel Collector, fan-out to both endpoints.
3. Decide on residency, defaults, and feature fit, not price. Both managed offerings are in the same ballpark.

Closing

SigNoz is a serious product run by a serious team. We're a different shape — managed-only, EU-default, with cardinality and alerting opinions baked in. Pick the one whose shape matches yours. The protocol-war pillar post is the broader category map; the OTel use-case page is our architectural detail.

If your shape is OSS-first or self-host-mandatory, SigNoz is the right answer. If your shape is managed-EU-with-cardinality-control, ours is.