Cloudflare — Engineering Performance
Avg. perf / dev / mo (ETV)
−8.1%
1.10 → 1.01
Active contributors
−3.6%
137.0 → 132.0
Growth
−5.7pp
36.5% → 30.8%
Fixes
−8.8pp
24.0% → 15.2%
Cloudflare vs. 500 OSS Performance Index
Per-engineer ETV for Cloudflare plotted against the pooled 500 OSS Performance Index. Both series are 90-day trailing rolling averages scaled to a 30-day month, so the curves sit on the same scale (ETV / dev / mo) and can be compared point-for-point. Latest reading: Cloudflare is 49% below the index (1.01 vs 1.97 ETV/dev/mo). Baseline gap was 34% below.
Monthly reports
Highlights
- *Wrangler CLI* was enhanced with a multi-profile authentication system, improving flexibility for developers to manage and switch between Cloudflare accounts [1cea1299 · emily-shen].
- A secure, opt-in HTTP fetch tool was added to the `@cloudflare/think` package, expanding *Think agents'* ability to interact with external APIs while ensuring strong SSRF defenses [de6a6951 · Sunil Pai].
- The *Think* core library was refactored to bundle `workers-ai-provider`, simplifying AI model configuration and providing a zero-boilerplate experience for `@cf` models [51ec433b · Sunil Pai].
- New capabilities were introduced to the *API schema* for audit logs, including product categories and filtering [1e8abab5 · cloudflare-schema-publisher[bot]], and for retrieving a page's *accessibility tree* [6b7f1c76 · cloudflare-schema-publisher[bot]].
- CI/CD reliability was improved by adding a release artifact backfill workflow [847883aa · Naresh] and implementing retry logic for `changeset version` on transient GitHub API failures [7c444b92 · Sunil Pai].
Observations
- Total output increased 40% compared to the 5-month average (133 current vs 95 average), indicating a highly active development period.
- *Maintenance* activity surged 83% compared to the 5-month average, reaching 67 points (current) vs 37 points (average). This included extensive updates to *OpenAPI schemas* (e.g., [000a8bac · cloudflare-schema-publisher[bot]], [5c28ad7a], [37abd721]), integration of new linting rules for deprecated APIs [f9c5b81b · Dario Piotrowicz], and hardening of browser test teardown processes (e.g., [16f78c66 · Sunil Pai], [59f7bb76]).
- The *Waste* score increased 35% compared to the 5-month average, reaching 32 points (current) vs 24 points (average). A significant portion of this waste stemmed from a large package versioning revert [75d8cb0e · Pete Bacon Darwin], which undid preparations for numerous *Wrangler* and *create-cloudflare* features.
- Several critical bug fixes were implemented, including resolving Durable Object OOM crash loops in the *chat recovery* system [1bbd9bca · Sunil Pai], fixing *Think* attachment metadata loss [62b90eba · Sunil Pai], and addressing duplicate resource bindings in *Wrangler* [6091d810 · James Anderson].
- A recurring pattern of updates and refactoring to the `openapi.yaml` schema was observed across multiple commits (e.g., [e61d7ec6], [bda160b7], [9c31b9c1], [1e8abab5], [83c11084]), indicating ongoing efforts to refine and expand *API definitions*.
- Test infrastructure received significant attention, with updates to replace deprecated Vitest utilities (e.g., [ddf3f4c9 · Dario Piotrowicz], [8eacfd18]) and the addition of regression tests for panic hook logging [fb9ce7c1 · Leo Blöcher], enhancing the robustness of the testing suite.
Repositories
Active repositories ranked by average performance per developer per month (over the last 90 days). The chart shows monthly performance composition — each repo as a stacked layer, with the top of the stack representing total org performance per month. Top 9 repos shown; the remainder is aggregated as “Other”.
| Repository | ||||
|---|---|---|---|---|
| agents | 7 | 167 | 7.9 | +1294%since Q2 2025 |
| sandbox-sdk | 5 | 26 | 1.7 | +348%since Q2 2025 |
| workers-sdk | 32 | 86 | 0.9 | +142%since Q2 2025 |
| workers-rs | 1 | 2 | 0.6 | +47%since Q2 2025 |
| lol-html | 1 | 2 | 0.6 | −5%since Q2 2025 |
| foundations | 1 | 1 | 0.5 | +2313959%since Q2 2025 |
| pingora | 6 | 8 | 0.5 | +114%since Q2 2025 |
| cloudflare-docs | 89 | 94 | 0.4 | −5%since Q2 2025 |
| cloudflared | 4 | 3 | 0.2 | −9%since Q2 2025 |
| containers | 2 | 1 | 0.2 | −80%since Q2 2025 |
| workerd | 30 | 10 | 0.1 | +16%since Q2 2025 |
| telescope | 2 | 1 | 0.1 | +396%since Q4 2025 |
Company total12 repositories | 132unique devs | 401ETV total | 1.01ETV / dev / mo | +85%since Q2 2025 |
| Performance (ETV) is the sum of every repository above. Active devs at the company level counts unique contributors across all repos, so a contributor working in multiple repos is counted once here but appears in each repo's row (the per-repo column will sum higher). ETV / dev / mo = Company ETV ÷ unique devs ÷ 3 mo. The "Since start" column compares each repo's Q1 2026 quarterly performance to the first quarter it had any activity — for repos that existed in Q2 2025 (when this index began), that's Q2 2025; for younger repos it's the quarter they actually started. The company row uses Q2 2025 as the baseline since the index itself began then. | ||||
Performance Growth vs Active Contributors
Engineering performance is outpacing team growth by 39.5×. Left axis shows total performance score, right axis shows active contributor count. The gap between curves represents productivity gains — more delivered per person, not just more people. Unit: Engineering Throughput Value (ETV).
Cost per Performance Unit
−43%
If performance per engineer grew 75%, each unit of engineering performance now costs approximately 43% less than at the baseline 90-day window (ending 2025-06-29). This is a directional estimate — the exact figure depends on fully-loaded engineer cost, but the direction is unambiguous.
Effective Capacity Added
+99 engineers
At today's productivity, the current 132-person team delivers the performance equivalent of 231 engineers at the baseline 90-day rolling window (ending 2025-06-29). That's roughly 99 engineers worth of capacity added through productivity gains, not hiring.
Performance Composition
Stacked bars show total complexity performance split into Growth (new value), Maintenance (sustaining systems), and Fixes (rework). The yellow line overlays performance per contributor — rising line means each engineer is delivering more, regardless of team size changes. Unit: Engineering Throughput Value (ETV).
CapEx vs OpEx
Monthly CapEx vs OpEx split. CapEx (capitalizable investment) is Growth — new features and capabilities. OpEx (operating expense) is Maintenance plus Fixes — keeping the lights on and reworking what's already shipped. The yellow line is the CapEx share, a quick read on how much of the month went into building new vs sustaining existing. Unit: Engineering Throughput Value (ETV).
Hours per Repository
Trailing 90-day window (64 working days). Org-level capacity is allocated to each repo by its share of org performance, then split CapEx / OpEx by that repo's own Growth vs Maintenance + Fixes mix.
| pingora | 67.3% | 32.7% |
| telescope | 52.9% | 47.1% |
| workerd | 48.6% | 51.4% |
| foundations | 48.0% | 52.0% |
| workers-rs | 48.0% | 52.0% |
| sandbox-sdk | 41.9% | 58.1% |
| workers-sdk | 33.8% | 66.2% |
| agents | 31.1% | 68.9% |
| cloudflare-docs | 19.6% | 80.4% |
| cloudflared | 11.2% | 88.8% |
| lol-html | 7.7% | 92.3% |
| containers | 6.4% | 93.6% |
| Total | 30.7% | 69.3% |
Fix Burden Distribution
Monthly rework volume broken down by who did it. Top contributors carry their named slice; everyone else is rolled into Others. Use this to spot whether fix work is concentrated on a small group (bus-factor risk) or distributed across the team.
Quarterly Summary
The raw numbers behind the charts: commits analyzed, active contributors, total performance, performance per developer, and the Growth / Maintenance / Fixes split for each quarter.
| Quarter | ||||||||
|---|---|---|---|---|---|---|---|---|
| Q2'25 | 2,510 | 130 | 227.99 | 0.6 | 31.3% | 57.1% | 11.6% | — |
| Q3'25 | 2,754 | 137 | 256.52 | 0.6 | 30.2% | 51.5% | 18.4% | +13% |
| Q4'25 | 2,410 | 130 | 253.18 | 0.6 | 27.4% | 51.8% | 20.8% | −1% |
| Q1'26 | 2,670 | 139 | 422.77 | 1 | 36.1% | 38.1% | 25.8% | +67% |
Top Contributors
Contributors ranked by performance per month (Growth + Maintenance + Fixes), over the last 90 days normalized to a 30-day calendar month.
The best way to measure AI efficiency
SampleA preview of the Navigara engine running on a sample organization. The numbers below are illustrative, not part of the OSS500 benchmark above.
Measure
Score every commit by depth
GitHub commits are weighted by what it took to write them, not by lines of code. The result is ETV per developer per month.
SourceGitHub
Spend
Tie ETV to cost
AI token bills and seat costs are pulled per team and divided by the ETV produced. The result is your true cost per unit of work.
SourceToken usage + finance
Map
Tie work to objectives
Each ETV is mapped to your Jira epics and labels, so you can see what's key-aligned, aligned, or unmapped capacity.
SourceJira
Performance
ETV delivered per developer / month
9.4ETV / dev / month
5.6 below target- Non-AI 5.8
- AI 3.6
AI Efficiency
AI spend per ETV unit delivered
$4.20
$0.60 over target- Cost / ETV $4.20
Objective Alignment
Share of work mapped to key objectives
51%
24 pts below target- Key-aligned 28%
- Aligned 23%