All nodes nominal

Your stack, from orbit.

Perigee watches every region, every service and every dependency from one view. You see the failure while it's still a flicker, not a postmortem.

Start tracking Watch an incident
Uptime
99.993%
Stations
19live
MTTD
19sec
Scroll
01 / Tracked by
Watching 41,208 endpoints across 19 regions
02Mission log

Incidents read like a flight recorder.

Every incident is a timeline, not a wall of red. Perigee writes the log while it's happening, in the order it happened, with the evidence already attached. By the time your on-call opens the laptop, the story is written.

Detection to page
19s
Blast radius named
2 of 19 regions
Raw telemetry kept
13 months
incident 4471 · ap-south-1 T+00:00:00
Duration 41m 06s Pages sent 1 Customers notified 0
03Instruments

Six instruments. One panel.

Nothing here is a widget you have to assemble. Perigee ships opinionated instruments, wired to each other on day one.

I-01

Orbit map

Every region on one sphere. Click a node and get its dependency tree, its last deploy, and the three checks that touch it. No dashboard archaeology.

I-02

Dependency arcs

Perigee maps what talks to what by watching real traffic, not a YAML file somebody wrote in 2023. When a shared path degrades, you see the blast radius before the pages start.

I-03

Ground stations

Checks run from 19 regions on real networks. If your CDN is fine in Frankfurt and broken in São Paulo, you'll know which one your customer is sitting on.

I-04

Black box

Every incident keeps 13 months of raw telemetry. Scrub back to any second, replay the exact request that broke, and hand the link to whoever wakes up next.

I-05

Synthetic orbits

Scripted user journeys that run every five seconds. Login, search, checkout. If the flow breaks, the alert names the step that broke, not the host it ran on.

I-06

Recovery signals

Perigee watches the fix land. It closes the incident when the metric actually recovers, not when somebody remembers to click resolve on Monday.

04Flight deck

The whole stack, one console.

Topology, latency and ground track share a single surface. Pick an instrument.

Legend
Nominal
Degraded
Observed path
Saturating
2026-07-15 · 04:22:06 UTC
Regions · 6 of 19
edge gateway auth-svc checkout search ledger-db cache POOL 96% OBSERVED TRAFFIC · LAST 60S · 7 SERVICES
Open incidents
1
ap-south-1 · 41m
p99 latency
58ms
↓ 12ms vs 1h ago
Error budget
97.4%
28d window
Checks / min
41,208
19 ground stations
24h now
05Pricing

Priced per region, not per panic.

No per-seat tax on the people who happen to be awake. Add your whole team on every tier.

Probe
$0/ forever

One region, one service, and enough rope to decide if you like it.

  • 3 ground stations
  • 20 monitors, 60s checks
  • 7 day replay
  • Email and Slack alerts
  • Unlimited seats
Launch a probe
Most picked
Station
$180/ month

The full instrument panel. Dependency arcs, on-call routing, 13 month black box.

  • All 19 ground stations
  • 500 monitors, 5s checks
  • Dependency arcs and blast radius
  • 13 month black box replay
  • On-call routing and runbooks
  • 99.99% uptime SLA
Start tracking
Constellation
Custom

For stacks that answer to auditors, and for regions that don't exist on a public map.

  • Everything in Station
  • Private ground stations in your VPC
  • SSO, SCIM, audit log export
  • 99.995% uptime SLA
  • A named engineer who knows your stack
Talk to us

All tiers include unlimited seats, unlimited alerts, and the status page.

Point it at your stack.

Perigee reads your first region in about four minutes. No agent to install, no sidecar, no YAML. Paste a URL and watch it light up.

Start tracking Read the build guide