About

Observability the size of your team.

Datadog-class visibility is priced for enterprises. Polartrace exists for everyone else: seed-to-Series-A teams running Node.js and Python services, who need traces, logs, metrics and alerts - and a bill they can explain.

Why we built it

Small teams get pushed into a bad choice: pay an observability bill that rivals payroll, or fly blind with scattered logs and a status page. The tooling itself isn't the hard part anymore - OpenTelemetry solved instrumentation. The hard part is a product priced and shaped for a team of five, not five hundred.

So Polartrace is deliberately narrow and deliberately cheap to run: two agents that install in two commands, six product surfaces that cover an incident end to end, alerting in the two channels startups actually use, and one subscription with two published meters. Nothing you have to configure for a week, nothing you have to negotiate to afford.

Principles

Commitments the code keeps

Not values on a wall - behaviors implemented in the product.

The bill is part of the product

Every overage rate is printed on the pricing page. Billable hosts are the 95th percentile of daily distinct hosts - a deploy spike never sets your bill. Your first overage invoice is waived, invoices under $1 are never charged, and you can set a spend cap that pauses ingest instead of billing you.

Never break the thing you monitor

The agents are built to fail safe: bounded queues that drop oldest first, exponential backoff that honors Retry-After, timers that never keep your process alive, and self-disable on misconfiguration. A broken agent costs you telemetry, never uptime.

Your data, minus your secrets

Passwords, tokens, authorization headers, cookies and card fields are replaced with [REDACTED] on your servers, before transmission. SQL string literals are scrubbed to placeholders.

Say what ships

The features page lists what exists today - Node.js and Python agents, six product surfaces, Slack and email alerts. Languages on the roadmap are labeled as roadmap. No aspirational feature grid.

Open where it matters

The agents that run inside your processes are MIT-licensed and public on GitHub, published on npm and PyPI. You can read every line that touches your traffic.

Built on OpenTelemetry

Instrumentation rides the OpenTelemetry SDKs and semantic conventions rather than a proprietary format - the industry's shared vocabulary for traces.

Open source

Read the code that runs in your app.

Both agents are MIT-licensed, public, and published to the registries you already trust.

Want to talk to the people building it? We answer our own support email.

Free plan: 5 GB/month - never billed

Start monitoring in minutes.

Two commands to your first trace. Start on the free plan - no credit card required.