Traceway: Self-Hosted MIT-Licensed Observability Stack (Deploy in 90 Seconds)

Published: 2026-05-13 • Reading: 12 min • 中文版

Traceway is an MIT-licensed, OpenTelemetry-native observability platform that combines logs, traces, metrics, session replay, exceptions, and AI observability into a single, self-hosted system. The headline number: you can deploy it in roughly 90 seconds with a single docker compose up -d command.

On Hacker News, the project hit 31 points with discussion centered on how it compares to existing open source observability tools like SigNoz and Grafana's Loki+Tempo+Mimir stack. The key differentiator: Traceway is MIT licensed without any "open core" restrictions — no BSL, no paid features hidden behind a license wall. Every feature is in the box.

If you're a small-to-medium team tired of SaaS observability bills or privacy-conscious about sending telemetry data to third-party clouds, Traceway is worth a serious look. Here's everything you need to know.

What Is Traceway?

Traceway is an observability backend designed from the ground up around OpenTelemetry (OTel). Point an OTLP exporter at it — any OTel SDK, any language — and telemetry starts flowing. No Collector, no glue code, no per-language vendor SDK. Just OTLP/HTTP ingest for traces, metrics, and logs.

The project describes itself as "the only tool you need to know what is happening and how to fix it." That's an ambitious claim, but the feature set backs it up:

MIT License: No Asterisks

The most notable decision in Traceway's design is the MIT license. This matters because observability tools have a complicated history with licensing in the open source world.

Popular alternatives often use restrictive licenses. Grafana's Loki and Tempo are AGPLv3, Elasticsearch switched from Apache 2.0 to SSPL/Elastic License, and several newer projects use Business Source License (BSL) or "open core" models where certain features require a paid subscription. HashiCorp's Terraform and Consul moved from MPL to BSL in 2023 — a pattern that has made teams wary of building on projects that might change their license terms later.

Traceway goes the opposite direction. The entire platform — including session replay, AI observability, multi-tenancy, and alerting — ships under MIT. There is no paid tier of self-hosted Traceway. The project monetizes through Traceway Cloud, a managed hosting service, but the code running that service is the same MIT-licensed code anyone can run themselves.

The comparison table from the Traceway README makes the positioning clear:

90-Second Deploy: How It Works

The most impressive part of Traceway is the setup experience. Here's the full deploy sequence:

git clone https://github.com/tracewayapp/traceway
cd traceway && docker compose up -d
# ✓ dashboard at http://localhost

That's it. The Docker Compose stack runs three services:

After starting, open http://localhost/register to create your first account, then point any OTel SDK at http://localhost/api/otel/v1/traces (or /metrics, /logs). Traces start appearing in the dashboard immediately.

For teams that want even less infrastructure, Traceway also supports an embedded mode — run the entire observability backend inside your Go process with SQLite instead of ClickHouse+PostgreSQL:

go get github.com/tracewayapp/traceway/backend

import tracewaybackend "github.com/tracewayapp/traceway/backend"

func main() {
  go tracewaybackend.Run(
    tracewaybackend.WithPort(8082),
    tracewaybackend.WithDefaultUser("admin@localhost.com", "admin"),
    tracewaybackend.WithDefaultProject("My App", "go", "dev-token"),
  )
}

This embedded mode is ideal for development, CI/CD pipelines, or single-instance deployments where running ClickHouse would be overkill.

Technology Stack

Traceway's architecture reflects sensible choices for a modern observability platform:

The Go backend was a deliberate choice — it means a single binary deployment with no runtime dependencies beyond the database. No JVM, no Node.js server process, no Python runtime. Combined with Docker Compose, this makes the "90-second deploy" claim genuinely achievable.

Client SDKs and Integrations

Traceway works with any OTel-compatible SDK, but also ships first-party integrations that include session replay:

Backend Integrations (Go ecosystem)

Full Stack / Frontend Integrations

Every integration ships traces, metrics, and logs over OTLP/HTTP. No proprietary SDK required. For frontend frameworks, session replay is included automatically.

Traceway vs. the Alternatives

Traceway enters a crowded field. Here's how it compares to the main alternatives teams actually consider:

vs. SigNoz

SigNoz is the closest open source competitor — also OTel-native, also self-hostable, with a similar feature set. Both use ClickHouse under the hood. The key differences: SigNoz's enterprise features are under a proprietary license (it follows an open-core model), while Traceway is fully MIT. Traceway also ships session replay and AI observability built-in, whereas SigNoz requires separate integrations for those capabilities.

vs. Grafana Stack (Loki + Tempo + Mimir + Grafana)

Grafana's ecosystem is proven at scale, but the operational complexity is real — you're managing at least four separate services plus the OTel Collector. Each component has its own configuration, retention policies, and scaling behavior. Traceway's single-service deployment is dramatically simpler for small-to-medium teams.

vs. Datadog / New Relic

The SaaS incumbents win on polish, breadth of integrations, and zero operational overhead. But they lose on cost — observability bills are a running joke in the industry for a reason. For a team of 10-50 engineers, Traceway self-hosted can save thousands of dollars per month while providing comparable visibility.

vs. OpenObserve

OpenObserve is another interesting open source option (AGPL-licensed) with similar capabilities. It claims significantly lower storage costs than Elasticsearch. Traceway's MIT license gives it a legal edge for teams sensitive to AGPL's virality provisions in commercial environments.

When Should You Use Traceway?

Traceway is a strong fit for these scenarios:

It's probably not the right choice for:

Getting Started with Traceway

The fastest way to try Traceway is the Docker Compose route mentioned above. For production, the documentation recommends:

Full walkthroughs are available at docs.tracewayapp.com.

Community and Roadmap

Traceway is built in the open, and the Discord community is active. The team uses it to discuss feature requests, help with setup and OTel wiring, share deployment stories, and give early access to upcoming features.

Contributions are welcome — pull requests get reviewed and merged. The project is at an early enough stage that community input genuinely shapes the roadmap. Missing a framework integration? You can open an issue or contribute one yourself.

Final Thoughts

Traceway isn't trying to beat Datadog at its own game. It's trying to make observability accessible — to teams that don't have the budget for enterprise SaaS, don't want to manage a zoo of Grafana components, or can't send their telemetry data to third-party clouds.

The MIT license is the real story here. In an industry where open source observability tools increasingly carry restrictive license terms, Traceway's commitment to permissive licensing removes a class of legal and operational concerns that teams have to think about when building on top of other tools.

90 seconds to deploy, no per-event billing, and the same code that runs the cloud service runs on your hardware. That's a compelling value proposition — and one that earned its 31 points on Hacker News for good reason.

Links:
GitHub Repository • Website • Documentation • Traceway Cloud • Discord Community • Hacker News Discussion