verify the deployment — formal confirmation in runbooks and checklists
toggle a feature flag — on/off switching; not "enable" or "activate"
0 / 5 completed
1 / 5
A developer clicks the "Deploy" button in their CI/CD interface to kick off the deployment process. Which phrase most naturally describes this action?
Trigger — the domain-specific CI/CD verb for initiating a pipeline:
Trigger is the standard verb in CI/CD vocabulary. A pipeline is "triggered" by an event — a push to a branch, a merged PR, a manual click, a scheduled cron, or a webhook. The word implies an event-driven initiation rather than a manual start.
Why not the others?
start — used colloquially and understood, but less precise; "start" implies a manual, direct action rather than an event-driven trigger
run — also informal but used ("run the pipeline", "run the build"); "trigger" is more idiomatic in written CI/CD documentation
launch — rarely used for pipelines; "launch" collocates with applications, rockets, and products
Trigger collocations in CI/CD:
trigger the pipeline / trigger a build / trigger a deployment / trigger a job
"The push to main triggered the release pipeline."
"You can trigger the deploy manually from the Actions tab."
In GitHub Actions: workflows are triggered by events defined in the on: block.
2 / 5
After initiating a deployment, a developer watches the dashboard to observe the deployment spreading gradually across servers. Which phrase is most natural in a DevOps context?
Monitor the rollout — the DevOps/SRE collocation for active deployment observation:
Monitor the rollout is the technical collocation used in DevOps and SRE contexts. A "rollout" is the progressive or gradual deployment of a new version to servers, containers, or users. "Monitor" implies active, sustained observation using metrics, dashboards, and alerts — not a one-time glance.
Why not the others?
watch — informal and used in conversation ("I'm watching the deploy"), but "monitor" is the professional term in runbooks and documentation
check — implies a one-time inspection; monitoring is continuous and ongoing during a rollout
follow — not natural in technical English for this context
Rollout vocabulary:
canary rollout — deploy to a small percentage of users first
In a runbook: "After triggering the deployment, monitor the rollout in Datadog for 15 minutes before closing the change window."
3 / 5
Immediately after a deployment completes, a developer opens their error monitoring tool to look at the rate of errors occurring in production. Which phrase is most natural?
Check error rates — the natural post-deployment verification collocation:
Check error rates is the natural collocation for the post-deployment step of looking at error monitoring dashboards (Sentry, Datadog, New Relic, CloudWatch) to verify the deployment has not introduced a regression.
Why not the others?
review error rates — "review" implies deeper, more structured analysis (a post-incident review, a weekly review); "check" is the right register for a quick, real-time verification step
measure error rates — "measure" describes what the monitoring tools do; the developer's action is to check/read what the tools have already measured
count error rates — wrong collocation; you count errors, but "rate" already implies a measured frequency, so "count rates" is redundant and unnatural
Standard post-deploy checklist: check error rates → check latency → check the logs → check downstream services.
4 / 5
The deployment has completed and the developer goes through a checklist to confirm that the application is running correctly and all systems are healthy. Which phrase is most precise?
Verify the deployment — the formal term used in runbooks and post-deploy checklists:
Verify the deployment is the formal, precise collocation used in deployment runbooks, change management processes, and post-deploy checklists. It implies systematic confirmation that the deployment completed successfully and the application behaves correctly.
Why not the others?
check the deployment — informal and understood, but "verify" carries more weight in formal documentation
test the deployment — implies running tests; smoke tests are part of verification but "test" has a more specific technical meaning
confirm the deployment — "confirm" collocates with bookings, approvals, and meetings; in tech it is used for confirming completion ("The deployment is confirmed complete") but "verify" is more appropriate for the active checking process
Deployment verification activities:
smoke tests — minimal tests to confirm critical paths work
health checks — /health endpoint returns 200
endpoint checks — verify key API responses
compare error rates before and after the deployment
Collocations: verify the deployment, verify the rollout, deployment verification, smoke test the deployment.
5 / 5
A team wants to enable a new checkout UI for 10% of users in production without redeploying the application. Which phrase correctly describes the action of changing the flag state?
Toggle a feature flag — the standard collocation for switching flag state:
Toggle a feature flag is the standard collocation because "toggle" precisely captures the on/off switching nature of feature flags. Feature flags (also called feature toggles or feature switches) allow teams to turn features on or off in production without a new deployment.
Why not the others?
switch a feature flag — a near-synonym, and understood, but "toggle" is more idiomatic and widely used in feature flag tool interfaces (LaunchDarkly, Flagsmith, Split)
enable a feature flag — describes turning a flag on only; "toggle" covers both enabling and disabling (the full lifecycle of managing flag state)
activate a feature flag — similar to "enable"; doesn't capture the reversible, on/off nature of toggling
Feature flag use cases:
canary releases — enable for 1% of users, monitor, then gradually increase
A/B testing — enable variant A for 50%, variant B for 50%
kill switch — immediately disable a feature without a deploy if something goes wrong
dark launching — enable in production but only for internal users