Continuous runtime tracing
Every request, job, and queue event is traced with sub-millisecond overhead — no sampling games, no blind spots between services.
Now in public beta
See every signal.
See every signal.
Block what matters.
Nimbus watches your services at runtime, learns what normal looks like for each endpoint, and blocks anomalies before they reach production — with traces your engineers can actually read.
SOC 2 Type II · 2,400+ engineers in the beta · one-command install
Capabilities
Observability and defense in a single pass
Most teams bolt a security tool onto an APM and reconcile two timelines at 3 a.m. Nimbus does both jobs from one agent.
Anomaly blocking with reasons
Nimbus profiles each endpoint for seven days, then denies out-of-profile traffic at the edge — every block tagged with a plain-English reason.
Triage from your terminal
Investigate without leaving the CLI you already live in: nimbus trace, nimbus block, nimbus explain.
How it works
From install to insight in three steps
No rules to write, no dashboards to babysit. Nimbus earns its keep in the first week.
01
Install the agent
One line in CI — or a single Helm chart — adds the Nimbus sidecar to every service. No code changes, no config files.
02
Learn the baseline
For seven days Nimbus profiles what normal means per endpoint: traffic shape, payloads, dependencies. Thresholds tune themselves.
03
Block with reasons
Out-of-profile requests are denied at the edge and logged with a full context trace, so on-call trusts the block instead of fighting it.
From the beta
Engineers who got their nights back
“We cut mean time-to-resolution from 40 minutes to six. Blocks arrive with a reason string attached, so on-call actually trusts them.”
“Nimbus replaced two tools and a spreadsheet of regex rules. Setup took one afternoon, and the traces read like our own logs.”
Pricing
Priced for teams that ship
Free while the beta runs. Upgrade when your pager agrees.
Solo
For side projects and kicking the tires.
$0per user / month
- 1 service, 24-hour trace retention
- Anomaly detection (log-only)
- Community support
Team
For production teams on call.
$29per user / month
- Unlimited services & environments
- 30-day retention & anomaly blocking
- Slack, PagerDuty & webhook alerts
- SSO via Google & GitHub
Enterprise
For regulated and self-hosted stacks.
Customannual contract
- Self-hosted or in-VPC control plane
- SSO, SCIM & audit-log export
- Dedicated support engineer
- 99.99% uptime SLA
FAQ
Questions we get at every demo
How does Nimbus learn what “normal” looks like?
During a seven-day profiling window the agent builds a per-endpoint baseline of traffic shape, payloads, and dependencies. No rules, no regex — the baseline comes from your real production behavior and re-validates on every deploy.
Can Nimbus run inside our own VPC?
Yes. On the Enterprise plan the control plane and trace store run fully self-hosted or inside your VPC. The agent is a single static binary, and no trace data ever leaves the boundary you define.
What happens when Nimbus blocks a request by mistake?
Every block ships with a full context trace and a reason string. One click adds a scoped allow rule, and the feedback folds back into the baseline — blocks are logged, reversible, and never silent.
Still have questions?
Put your runtime on watch
Join the public beta and get the full Team plan free until 1.0 — one command to install, and your feedback shapes the roadmap.
No credit card · SOC 2 Type II · Cancel anytime