Budgets in the repo
A p99 target next to the endpoint, reviewed like any other change.
01Latency budgets as code
Declare a p99 in the repo. The build fails when it slips. The dashboard stops lying.
Lumi0 gives your team one place to latency budgets as code — with limits that are visible before they are hit.
No card required. 212→88 ms p99 in six weeks.
Overview
Runs
Policies
Traces
Settings
Latencylive
212→88 msp99 in six weeks
0silent regressions since
Budgets in the repook
Build fails on regressionok
Attribution to the queryok
Teams shipping with Lumi0
NorthwindKestrelRuangPolarTokoLumenBakauVela
212→88 msp99 in six weeks
0silent regressions since
4lines to declare a budget
Inside the product
One screen, the whole picture.
Overview
Runs
Policies
Traces
Settings
Latencylive
212→88 msp99 in six weeks
0silent regressions since
Budgets in the repook
Build fails on regressionok
Attribution to the queryok
What you get
Three things Lumi0 does properly.
Build fails on regression
No more “we will look at it next sprint” — the pipeline says no.
02Attribution to the query
The slow span names the exact SQL, not just the controller.
03“The first week the build failed on a query nobody had looked at in a year. It was 40% of p99.”
Ren T. — product engineer, Tokyo
Pricing
Start free. Pay when it earns it.
Questions
The things people ask first.
Do you need traces?+
No. Histograms from your existing metrics are enough to enforce a budget.
What counts as a regression?+
You define it: absolute p99, or a delta against the rolling baseline.
Can I gate only some routes?+
Yes, per route and per environment.
Lumi0 is open for design partners.
Small team, direct access to the people who build it. Bring a real workload.
Request access