Practice & Reference · Flashcards
Flashcards
38 cards drawn from every module in this course, front to back. The front holds a question, the back holds the answer — click or tap a card to flip it, and search to jump straight to a topic.
☺ Explain it like I'm 10
Rereading your notes feels like studying because the words look familiar, but recognizing a fact and being able to produce it from memory on your own are two different skills — like humming along to a song versus singing it with no music playing. Flipping a card forces the harder skill: you have to attempt the answer before you're allowed to check it. Do a few cards most days instead of cramming the whole deck once, and the same facts take less effort to recall each time you come back to them.
Foundations
What is DevOps, in one sentence — and what is it not?
Tap / Enter to flip →
Answer
A set of cross-functional practices that closes the gap between the people who write software and the people who run it, aligning their incentives — not a job title, not a tool, and not a reorg that creates a third team between Dev and Ops.
Foundations
What do the five letters of CALMS stand for, and who added the L?
Tap / Enter to flip →
Answer
Culture, Automation, Lean, Measurement, Sharing. The original CAMS (Debois and Shafer) became CALMS when Jez Humble and Damon Edwards added Lean around 2010.
Foundations
Which 2009 conference talk is credited as DevOps' founding moment, and who coined the term itself?
Tap / Enter to flip →
Answer
John Allspaw and Paul Hammond's "10+ Deploys Per Day: Dev and Ops Cooperation at Flickr" at O'Reilly Velocity, June 2009. Patrick Debois coined "DevOps" a few months later, naming DevOpsDays Ghent (October 2009) after it.
Foundations
State Conway's Law.
Tap / Enter to flip →
Answer
Organizations which design systems are constrained to produce designs that are copies of the communication structures of those organizations — communication paths become API boundaries whether or not anyone planned it that way. (Melvin Conway, 1967)
Foundations
What is the reverse Conway maneuver?
Tap / Enter to flip →
Answer
Deliberately restructuring teams around the architecture you want, and letting Conway's Law do the rest of the work — instead of letting an existing org chart accidentally dictate the system's shape.
Foundations
Name Westrum's three organizational culture types.
Tap / Enter to flip →
Answer
Pathological (power-oriented — information hoarded, failure triggers blame), bureaucratic (rule-oriented — information moves through official channels, failure triggers a search for the violated policy), and generative (performance-oriented — information flows freely, failure triggers inquiry into what happened).
Foundations
List the eight stages of the DevOps lifecycle loop, in order.
Tap / Enter to flip →
Answer
Plan, Code, Build, Test, Release, Deploy, Operate, Monitor.
Foundations
Which single arrow in the DevOps lifecycle diagram is what actually makes it a loop, and what does it carry?
Tap / Enter to flip →
Answer
Monitor feeding back into Plan — it carries production telemetry (errors, latency, incident patterns) as a direct input to backlog prioritization, not just a dashboard someone might glance at.
Foundations
What does "you build it, you run it" mean, and who's credited with the phrase?
Tap / Enter to flip →
Answer
The team that writes a service also deploys, monitors, and carries the pager for it in production. Werner Vogels, Amazon's CTO, described Amazon's service-oriented model this way.
Delivery Pipeline
What's the core difference in branch lifetime between trunk-based development, GitHub Flow, and GitFlow?
Tap / Enter to flip →
Answer
Trunk-based branches live hours; GitHub Flow branches live roughly a day or two behind a pull request; GitFlow keeps long-lived develop, release/*, and hotfix/* branches around a formal release cycle.
Delivery Pipeline
What does a feature flag let a team do that a deploy alone can't?
Tap / Enter to flip →
Answer
Decouple deploy from release — ship code to production switched off, then release it later with a config change instead of a new deploy, and roll back instantly by flipping the flag back off.
Delivery Pipeline
Define Continuous Integration, Continuous Delivery, and Continuous Deployment precisely.
Tap / Enter to flip →
Answer
CI: merging small changes to trunk frequently, with an automatic build and test on every merge. Continuous Delivery: every passing change is verified and proven deployable, but a human still decides when to release. Continuous Deployment: that human gate is removed — every passing change ships to production automatically.
Delivery Pipeline
True or false: all continuous deployment is continuous delivery, but not all continuous delivery is continuous deployment.
Tap / Enter to flip →
Answer
True. Delivery is the necessary foundation — the codebase is always releasable; deployment is delivery with the release button wired to the pipeline instead of to a person.
Delivery Pipeline
Under the fast-feedback principle, why does lint run before a security or dependency scan?
Tap / Enter to flip →
Answer
Lint takes seconds and needs no built artifact; a security scan takes longer and needs an artifact to exist first. Cheapest, fastest checks run first so a broken change fails before the pipeline spends time on expensive stages.
Delivery Pipeline
What does "build once, promote everywhere" forbid?
Tap / Enter to flip →
Answer
Rebuilding the artifact from source at any stage after the first. The exact same digest that passed staging must be the one deployed to production — only tags move between environments, never the underlying bytes.
Delivery Pipeline
What's the difference between a mutable image tag and a digest?
Tap / Enter to flip →
Answer
A tag like
:latest is a pointer that can be repointed to a different image later; a digest (sha256:...) is a content-addressed hash that pins to one specific, unchangeable image forever.Delivery Pipeline
Rolling, blue-green, and canary deployment — what's the core trade-off each makes?
Tap / Enter to flip →
Answer
Rolling replaces instances gradually against one environment: cheapest, but rollback is slow. Blue-green cuts 100% of traffic at once between two full environments: near-instant rollback, double infrastructure cost. Canary shifts a small percentage of traffic gradually with metric gates: caps blast radius, needs strong real-time metrics.
Delivery Pipeline
How does a canary release differ from an A/B test, even though both route a percentage of traffic to a variant?
Tap / Enter to flip →
Answer
A canary is a reliability signal — does the new build error and perform as well as the old one, run by platform/SRE, torn down once rollout completes. An A/B test is a product/business experiment — does variant B convert or retain better, run by product/growth, kept running for a fixed sample window.
Infrastructure & Operations
What's the structural difference between declarative and imperative IaC?
Tap / Enter to flip →
Answer
Imperative specifies the sequence of steps to reach a result — a script of API calls. Declarative specifies the desired end state and lets the tool compute the diff and the create/update/delete calls needed to get there.
Infrastructure & Operations
What does it mean for an IaC apply to be idempotent?
Tap / Enter to flip →
Answer
Running apply again against unchanged configuration produces zero changes, because current state already matches desired state — the property that makes it safe to re-run after a failed apply or a flaky CI retry.
Infrastructure & Operations
What is infrastructure drift?
Tap / Enter to flip →
Answer
The gap that opens when real infrastructure changes without going through the IaC tool — a manual console edit, another automation acting outside it. State now disagrees with reality, so the next plan either silently reverts the fix or needs an explicit reconcile.
Infrastructure & Operations
What problem does IaC state locking solve?
Tap / Enter to flip →
Answer
It prevents two applies from racing against the same state at once. Without it, each apply can compute a plan against a snapshot that's already stale by the time it executes, corrupting the state file or issuing conflicting API calls.
Infrastructure & Operations
Where does infrastructure as code stop and configuration management start?
Tap / Enter to flip →
Answer
IaC provisions resources that don't yet exist — Terraform creates a VM. Configuration management configures software on a resource that already exists — Ansible installs and configures nginx on that VM.
Infrastructure & Operations
Push-based (Ansible) vs. pull-based (Puppet, Chef) configuration management — what's the core trade-off?
Tap / Enter to flip →
Answer
Push: a controller triggers runs on demand, simple to reason about, but doesn't self-heal between runs. Pull: an agent reconciles on its own schedule (Puppet's default is every 30 minutes), self-healing drift automatically, at the cost of running and securing an agent and server on every host.
Infrastructure & Operations
Why does a container start in milliseconds while a VM typically takes tens of seconds to boot?
Tap / Enter to flip →
Answer
A container shares the host's existing kernel and is isolated with namespaces and cgroups, so starting one is close to launching a process. A VM boots an entire separate guest operating system, kernel included, on virtualized hardware first.
Infrastructure & Operations
What do Kubernetes' Pod, Deployment, and Service objects each do?
Tap / Enter to flip →
Answer
A Pod is the smallest deployable unit — one or more containers sharing a network namespace. A Deployment manages a set of Pod replicas, driving self-healing and rolling updates. A Service gives that changing set of Pods one stable network identity and DNS name.
Infrastructure & Operations
Name Google's four golden signals for monitoring any service.
Tap / Enter to flip →
Answer
Latency (how long requests take, reported as a distribution, not an average), traffic (demand on the service), errors (rate of failed requests), and saturation (how full the service is relative to its limit — a leading indicator).
Infrastructure & Operations
What's the precise difference between monitoring and observability?
Tap / Enter to flip →
Answer
Monitoring watches a predetermined set of signals for predetermined failure modes. Observability is having enough raw, high-cardinality telemetry to answer novel questions about failure modes nobody predicted in advance.
Infrastructure & Operations
Beyond the engineer actively debugging, what roles typically staff an incident response?
Tap / Enter to flip →
Answer
An incident commander coordinates the response and makes calls without necessarily touching the fix themselves; a communications lead keeps stakeholders updated; a scribe logs the timeline — separating coordination from hands-on-keyboard work so the person fixing the issue isn't also managing the room.
Practice & Reference
Name the four DORA metrics.
Tap / Enter to flip →
Answer
Deployment frequency, lead time for changes, change failure rate, and time to restore service (MTTR).
Practice & Reference
Which two DORA metrics measure throughput, and which two measure stability?
Tap / Enter to flip →
Answer
Throughput: deployment frequency and lead time for changes. Stability: change failure rate and time to restore service.
Practice & Reference
What are the Elite-tier DORA benchmarks for deployment frequency and change failure rate?
Tap / Enter to flip →
Answer
Deployment frequency: on-demand, multiple deploys per day. Change failure rate: 0–15%.
Practice & Reference
What are the Elite-tier DORA benchmarks for lead time for changes and time to restore service?
Tap / Enter to flip →
Answer
Both under one hour.
Practice & Reference
What's the counter-intuitive core finding of the DORA research?
Tap / Enter to flip →
Answer
Elite performers don't trade stability for speed — they deploy far more often than low performers while also failing less and recovering faster, because both outcomes come from the same underlying practices: small batches, trunk-based development, automated testing.
Practice & Reference
What's the formula for change failure rate?
Tap / Enter to flip →
Answer
Failed deploys divided by total deploys over a period, expressed as a percentage. A deployment counts as "failed" if it causes a degraded service requiring a hotfix, rollback, or patch.
Practice & Reference
Who founded the DORA research program, and what book published its underlying model?
Tap / Enter to flip →
Answer
Nicole Forsgren, Jez Humble, and Gene Kim founded DORA (later acquired by Google); the statistical model behind the four metrics was published in their 2018 book Accelerate.
Practice & Reference
State Goodhart's Law and how it applies to gaming deployment frequency.
Tap / Enter to flip →
Answer
"When a measure becomes a target, it ceases to be a good measure." A team can split one meaningful release into ten trivial "deploys" to inflate the count on a dashboard without shipping more real value or improving lead time or failure rate.
Practice & Reference
Match each stage to a representative tool: CI/CD orchestration, infrastructure as code, configuration management, artifact registry.
Tap / Enter to flip →
Answer
CI/CD orchestration — GitHub Actions or Jenkins. Infrastructure as code — Terraform. Configuration management — Ansible. Artifact registry — JFrog Artifactory, Docker Hub, or GitHub Container Registry.