O glossário de métricas de engenharia.
Definições diretas das métricas que líderes de engenharia realmente reportam. DORA, flow, saúde de sprint e developer experience, com benchmarks, fórmulas e onde cada uma se encaixa no DevStats.
Comece pelas métricas mais buscadas.
Deployment frequency
Successful production deployments in a given period — daily, weekly, or on-demand.
Ler definição 02Change failure rate
Share of deployments that trigger a hotfix, rollback, or incident response.
Ler definição 03Lead time for changes
Time from first commit to production. Elite teams ship in under one hour.
Ler definição 04Mean time to detect
Gap between failure start and detection. An observability signal, not a recovery one.
Ler definiçãoMétricas DORA
Os quatro sinais definidos pelo DevOps Research & Assessment, mais as métricas de incidentes ao lado deles. Velocidade e estabilidade do seu pipeline de entrega.
Deployment frequency
Successful production deployments in a given period — daily, weekly, or on-demand.
Mais buscadaChange failure rate
Share of deployments that trigger a hotfix, rollback, or incident response.
Mais buscadaLead time for changes
Time from first commit to production. Elite teams ship in under one hour.
Mean time to recovery
Total downtime ÷ number of incidents. Lower is better; under an hour is elite.
Mais buscadaMean time to detect
Gap between failure start and detection. An observability signal, not a recovery one.
Mean time between failures
Average healthy interval between consecutive incidents on a service.
Mean time to acknowledge
Time between alert fire and a human on-call accepting the page.
Incident response time
End-to-end time from page acknowledgement to resolution path picked.
Accelerate metrics
The original four-metric framework introduced in the Accelerate research.
Software delivery metrics
Umbrella term for the throughput + stability metrics across the SDLC.
Change lead time
Alternate name for lead time for changes used in some incident frameworks.
Release cadence
The planned rhythm at which your team commits to ship to users.
Time to restore service
Synonym for MTTR — total time to return service to a healthy state.
Reliability metric
Generic family that includes MTBF, MTTR, error budget, and SLO burn.
Deployment success rate
Inverse of change failure rate, measured at the pipeline stage.
Rollback rate
Percentage of deployments that get reverted within a defined window.
Deployment cadence
How regularly successful production deployments occur over time.
Métricas de Engenharia
Throughput, flow, cycle time e os padrões dentro do ciclo de vida do pull request. O sinal do dia a dia de como o trabalho avança.
Cycle time
Active work start to delivery. Often confused with lead time — measures different spans.
Burndown chart
Remaining sprint work plotted against time. Surfaces commitment vs. velocity gaps.
Cumulative flow diagram
Stacked area showing WIP across pipeline stages. Reveals bottlenecks visually.
Sprint velocity
Story points (or count) a squad ships per sprint, used for forward planning.
WIP limits
Caps on how many items can sit in a given stage at once. Forces flow.
Flow metrics
Family covering velocity, cycle time, WIP, and aging — the Kanban-flow set.
Engineering productivity
Composite signal for how much value engineering delivers per unit of effort.
Development cycle time
Subset of cycle time covering the active coding + review stages only.
Flow efficiency
Active time ÷ total elapsed time across a PR or ticket lifecycle.
Code churn
Lines changed shortly after being written. A signal of rework or churn pressure.
Scope creep
Tickets or work added mid-sprint that were not in the original commitment.
Escaped defects
Bugs that reach production. A trailing indicator of testing and review rigor.
PR cycle time
Time from PR open to merge. The clearest single signal of review health.
Code review metrics
Review wait, review time, reviewer distribution — collaboration signals.
Coding time
First commit to PR ready-for-review. Stage 1 of the PR cycle.
Velocity metric
Generic shipping rate signal — story points, PRs, or features per sprint.
Deploy time
Final stage of PR cycle — merge to production deploy completing.
Software engineering productivity
O que significa produtividade em engenharia de software, como medir sem métricas de vaidade e como melhorar a entrega…
PR review time
Saiba o que é PR review time, como medir, os benchmarks e como reduzir a espera que desacelera os seus pull requests.
Software cycle time
Alternate phrasing of PR cycle time used in some delivery frameworks.
Sprint burndown
Burndown specialised to a single sprint cycle. Shows commitment risk early.
Context switching
Frequency a player jumps between concurrent issues, repos, or contexts.
PR size
Lines of code changed per pull request. The strongest predictor of review time.
Rework rate
Share of code modified again within a short window of being merged.
Unplanned work
Effort spent on items that did not exist at sprint planning. KTLO territory.
Commit frequency
Rate of commits per player per day or week. A raw activity signal.
Throughput metric
PRs, issues, or features completed in a period. Companion to cycle time.
Team throughput
Throughput aggregated at the squad level over weekly / monthly windows.
Delivery cadence
How consistently the squad ships work — variance matters as much as mean.
Sprint predictability
How close planned vs. delivered scope tracks across recent sprints.
Planning accuracy
Delivered work ÷ committed work. A trust signal with product partners.
Carryover work
Items committed in one sprint but completed in a later one.
Sprint completion rate
Entenda o sprint completion rate, como medir, os benchmarks e como entregar mais do que o seu squad se compromete a f…
Merge time
O que é merge time, como medir, os benchmarks e como reduzir a espera entre a aprovação e o merge de um pull request.
Review turnaround time
Saiba o que é review turnaround time, como medir, os benchmarks e como acelerar a primeira resposta nos pull requests.
Framework SPACE
Satisfação, Performance, Atividade, Comunicação, Eficiência. A lente de developer experience que acompanha DORA + flow.
Engineering allocation
Where dev time goes: roadmap, KTLO, bugs, hotfix. A reporting metric.
Mais buscadaDeveloper experience
Friction, tooling, and feedback loops players hit doing their day-to-day work.
Developer velocity
Composite of throughput and cycle time, framed as an experience metric.
Engineering velocity
Org-level shipping speed — usually throughput per engineer over time.
Engineering efficiency
How much delivered work translates to outcomes vs. activity-for-activity.
Developer burnout
Sustained after-hours activity, weekend commits, and rising rework rate.
Developer satisfaction
Self-reported NPS-style signal on tooling, ceremonies, and review culture.
Developer wellbeing
Pace sustainability + cognitive load patterns visible in activity data.
Knowledge silos
Code areas reviewed or modified by only one or two people.
Developer efficiency
Output per unit of effort, with effort proxied by active hours or context switches.
After hours work
Activity outside core hours — a leading indicator of unsustainable pace.