v4.2 · released · lower tail latencies across 21 regions

Observe everything.
Ship anyway.

Pulse is a fictional observability tool built as a portfolio template. Traces, logs, and metrics sit on one timeline, so the question “what changed?” always has a one-click answer — even on the worst nights.

See the stack
pulse · prod-eu · traces
● traceslogsmetrics
latency · p50
24.8ms
Δ -12.4% vs. 1h
spans · checkout flow · 5s
api.checkout
48ms
db.write
12ms
cache.get
3ms
auth.verify
19ms
api.search
112ms
db.read
8ms
p50 latency
21ms
global median
events/sec
4.8M
ingested yesterday
regions
21
live, bgp-announced
uptime
99.99
rolling 90-day
// why teams pick pulse

One timeline.
Every signal.

01 · trace

One timeline, every signal

Traces, logs, and metrics share a single cursor. Drag to compare, click to pin.

02 · query

A language you already think in

PulseQL reads like speech. If you can describe the incident, you can query it.

03 · alert

Alerts that respect sleep

Routing by team, by hour, by signal noise. Your 3am self will thank you.

in use at fictional teams of every size
◆ HalcyonBrass & Co.KODAMA△ northline/near/fieldpolaris•
// from the field
We replaced four dashboards with one Pulse timeline. The first time we found a bad deploy in forty seconds, the room went quiet.
JL
Jiho Lee
Staff SRE · Halcyon
// ready when you are

Point a collector at us.
See the graph fill.

$ install
curl -sSL pulse.dev/install | sh
see pricing →