liveBook · Manning

Headroom

As explained in Fast by Default

Budget headroom is the remaining margin between current resource or performance usage and an applicable budget limit.

Headroom is the remaining margin between current resource or performance usage and an applicable budget limit. It shows how much capacity remains before a change requires cleanup, negotiation, or an exception. section 6.3.3 section 6.5.2

Use headroom to make budget pressure visible and actionable:

  • Allocate it explicitly across teams, features, and reserve pools instead of allowing shared capacity to disappear silently. section 6.3.3
  • Set enforcement thresholds below user-facing goals, preserving a buffer that catches degradation before users notice it. section 6.4
  • Track it in a shared budget document so engineers can see remaining capacity and follow an explicit exception process. section 6.5.2
  • Treat shrinking headroom as an early warning: initiate cleanup below 10% rather than waiting for the budget to break. section 6.4

Search API pressure

A search API with a 150ms budget and only 8ms of headroom has little room for the next feature touching search. The team can optimize existing work, negotiate a trade-off, or request an exception with documented conditions. section 6.5.2

Figure 6.2 A budget allocation dashboard showing each team's current usage against their allocation. Core is nearly at capacity, while Analytics and Reserve have room for growth.

Use the allocation dashboard to compare each team's current usage with its allocation and identify where capacity is nearly exhausted or still available. section 6.3.3

Table 6.2 Starting point thresholds for common performance budgets
Platform
Resource
Aggressive
Moderate
Relaxed
Web
JavaScript bundle (compressed)
< 150KB
< 250KB
< 400KB
Web
Total page weight
< 1MB
< 2MB
< 4MB
Backend
API response time (p95)
< 100ms
< 200ms
< 500ms
Backend
Query time (p95)
< 20ms
< 50ms
< 100ms
Mobile
Cold start time
< 2s
< 3s
< 5s
Mobile
App size
< 30MB
< 50MB
< 100MB
Mobile
Memory (typical use)
< 100MB
< 200MB
< 400MB
Desktop
Launch time
< 2s
< 4s
< 8s
Desktop
Idle memory
< 300MB
< 500MB
< 1GB

Use the threshold table to establish starting budgets for bundles, page weight, API latency, startup, memory, and launch time. section 6.4

How Fast by Default covers it

How deep it goes
Defines the metric and applies it to allocation, threshold buffers, growth planning, and exception documentation.
Examples
Uses team allocation dashboards, API latency, JavaScript bundle size, mobile startup, app size, and desktop memory examples.
What it assumes
Assumes a measured baseline, an applicable performance goal, and a defined budget or hard limit.
What you can do after
Operate performance budgets by measuring remaining capacity, allocating it explicitly, and escalating shrinking margins before limits are breached.
How it connects
Connects performance budgets with performance goals, resource allocation, exception processes, and platform-specific metrics.

Also explained in

Compare how 4 books explain headroom →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Fast by Default. The section links point to the parts of the book it draws on.