Performance Budgets as Engineering Discipline

Performance budgets ratchet down or hold. Without them, performance degrades quietly with every release.

Why budgets prevent drift

Performance regressions are gradual; nobody notices any single one; cumulatively the site is slow a year later.

Budgets catch each regression at PR time.

Four budget categories

• 1. Bundle size.
• 2. p99 endpoint latency.
• 3. Memory per request.
• 4. CPU per request.

Per-budget threshold

Bundle size: ratchet down; never let increase without justification.

p99 latency: hold within 5% of last release.

Memory + CPU: hold within 10% of baseline.

Each PR’s impact measured.

CI integration

CI step: load test + assertion. Bundle-size analyzer + assertion. Fail PR if budget exceeded without ‘allow-regression’ tag.

Allow-regression with justification + dated cleanup ticket.

Antipatterns

• No budgets. Drift wins.
• Budgets without enforcement. Decoration.
• Allow-regression as default. Defeats the discipline.

What to do this week

Three moves. (1) Apply this pattern to your slowest production endpoint. (2) Measure p99 before/after. (3) Document the win and ship the runbook so the team can reproduce.