Developer Tool
Tracer
A Postgres intelligence tool designed for engineers who run everything themselves.
Client
Tracer
YEAR
Tracer started as a question rather than a client brief.
What would Postgres performance tooling look like if the interface was taken as seriously as the engineering? Solo developers and small teams run multiple databases simultaneously with no single view of what's slow, what's getting worse, or why. Existing tools are either enterprise APMs with six-figure contracts, or raw logs that need a DBA to interpret. Tracer is the middle a precision instrument for people who are both the engineer and the ops team.
The heatmap leads, not numbers. A full year of performance history as a colour-coded grid the May incident cluster visible before you've read a single label.
Status badges are calculations. A comment mid-project cut to the point: "Healthy/Warning/Critical only works if the thresholds are honest." One threshold number now cascades to every badge, dot, and alert in the product. The live preview shows your queries redistribute in real time as you drag the slider.
One interaction pattern, two entry points. The slide-in detail panel works identically whether triggered from the Query Ledger or the sidebar. Learn it once, use it everywhere.
17 screens, a full design system. The social response confirmed the audience immediately with engineers, debating threshold logic, asking where they could sign up. Several client enquiries followed.
Let's work together
Building a plugin, a platform, or something that doesn't fit a category yet, if your users can't afford a broken workflow, let's build a better one.
