Cloud Computing DevOps, Automation, and Operations Flashcards

24 sample cards from the DevOps, Automation, and Operations deck

Sampled from the end of the deck, so these are different cards from the ones shown on the syllabus page.

1. What are the 'three pillars of observability'?

   Metrics, logs, and traces. Metrics are aggregated numeric measurements over time; logs are timestamped discrete event records; traces track a single request's path across distributed services.

2. How do metrics, logs, and traces differ in structure and use?

   Metrics are low-overhead numeric time series good for dashboards/alerts. Logs are detailed, high-cardinality event text used for debugging specifics. Traces show causal, end-to-end latency of a request across microservices to locate bottlenecks.

3. What is distributed tracing, and what is a span?

   Distributed tracing follows a request as it propagates through multiple services. A span is a single named, timed operation (one unit of work) within a trace; spans nest to form the complete trace with a shared trace ID.

4. Define SLI, SLO, and SLA.

   An SLI (Service Level Indicator) is a measured metric of service behavior (e.g., latency, availability). An SLO (Service Level Objective) is the internal target for that SLI (e.g., 99.9% availability). An SLA (Service Level Agreement) is a customer-facing contract with consequences (e.g., refunds) if the objective is missed.

5. How is a typical availability SLI expressed as a formula?

   $$\text{Availability} = \frac{\text{good events}}{\text{total valid events}} \times 100\%$$ For example, $\frac{\text{successful requests}}{\text{total requests}}$ over a measurement window.

6. What is the relationship in strictness between an SLA, SLO, and SLI target?

   The SLO is set stricter than the SLA to provide a safety buffer, i.e., $\text{SLO} \geq \text{SLA}$. The SLI is the actual measured value you compare against the SLO. Teams alert on the SLO before the SLA is ever breached.

7. Define an error budget and give its formula.

   An error budget is the maximum allowable amount of unreliability over a period, i.e., the room between perfection and the SLO: $$\text{Error Budget} = 100\% - \text{SLO}$$ For a 99.9% SLO, the error budget is 0.1% of requests/time.

8. For a 99.9% availability SLO, roughly how much downtime is allowed per month?

   About $0.1\%$ of the time. Over a 30-day month ($43{,}200$ minutes), that is $\approx 43.2$ minutes of allowed downtime.

9. What is the purpose of an error budget in SRE practice?

   It balances reliability against feature velocity. As long as budget remains, teams can ship freely. When the budget is exhausted, releases are frozen and effort shifts to reliability work—turning reliability into a shared, data-driven decision.

10. What is the 'burn rate' of an error budget?

    The rate at which the error budget is being consumed relative to the SLO window. A burn rate of $1$ exhausts the budget exactly at the end of the window; a burn rate of $10$ exhausts it $10\times$ faster, triggering high-urgency alerts.

11. What is incident management?

    The structured process of detecting, responding to, mitigating, and resolving unplanned service disruptions to restore normal operation as quickly as possible, while communicating status and later learning from the event.

12. What is the role of an Incident Commander (IC)?

    A single person who owns coordination of an incident response—delegating tasks, tracking status, and communicating—without necessarily doing the hands-on fixing. The IC ensures clear ownership and avoids chaos during major incidents.

13. What is a blameless postmortem?

    A retrospective written after an incident that focuses on systemic causes and improvements rather than blaming individuals. It documents timeline, impact, root cause, and action items, encouraging honest disclosure to prevent recurrence.

14. Define MTTR, MTTD, and MTBF.

    MTTR = Mean Time To Recover/Repair (average time to restore service). MTTD = Mean Time To Detect (average time to notice an incident). MTBF = Mean Time Between Failures (average uptime between incidents): $$\text{MTBF} = \frac{\text{total operational time}}{\text{number of failures}}$$

15. What is on-call rotation and why is it used?

    A schedule assigning engineers responsibility for responding to production alerts during specific time windows. Rotations distribute the operational burden, ensure 24/7 coverage, and typically include primary and secondary (escalation) responders.

16. What is GitOps?

    An operational model where Git is the single source of truth for both application and infrastructure declarative configuration. A controller continuously reconciles the live cluster state to match the desired state committed in Git; changes are made via pull requests.

17. What are the core principles of GitOps?

    1) The entire system is described declaratively; 2) The desired state is versioned in Git; 3) Approved changes are automatically applied; 4) Software agents continuously reconcile actual state to desired state and can self-heal/alert on drift.

18. What is ArgoCD?

    A declarative, Kubernetes-native GitOps continuous delivery tool (a CNCF project). It runs as a controller inside the cluster, monitors Git repositories, and syncs the cluster to match the manifests, providing a UI showing sync and health status.

19. What is Flux (Flux CD)?

    A CNCF GitOps toolkit for Kubernetes that continuously reconciles cluster state from Git sources. It is composed of controllers (source, kustomize, helm, notification) and is often more CLI/GitOps-native and lightweight than ArgoCD.

20. Compare ArgoCD and Flux at a high level.

    Both are CNCF GitOps tools for Kubernetes. ArgoCD is application-centric with a rich web UI and built-in RBAC/SSO, using an Application CRD. Flux is a modular set of controllers, more Git-native and lightweight, integrating tightly with kubectl/Helm and lacking a first-party UI by default.

21. What is configuration drift?

    Divergence between the actual running state of infrastructure and the desired state declared in code/Git, caused by manual changes, hotfixes, or external processes making out-of-band modifications.

22. What is reconciliation in the context of GitOps and controllers?

    The continuous control-loop process where an agent compares actual state to the declared desired state and takes corrective action to converge them. This is the same converge loop Kubernetes controllers use to maintain desired replica counts, etc.

23. What is cloud cost monitoring and what are budgets/alerts?

    Cost monitoring is tracking and analyzing cloud spend over time (e.g., AWS Cost Explorer). Budgets set spending thresholds that trigger alerts (or actions) when actual or forecasted cost exceeds a defined limit, preventing surprise bills.

24. What are rightsizing and resource tagging in cost management?

    Rightsizing means matching provisioned resource capacity (instance type, size) to actual utilization to eliminate over-provisioning waste. Resource tagging assigns key-value metadata (e.g., env, team, cost-center) to resources so costs can be allocated, filtered, and attributed accurately.

What this deck covers

DevOps, Automation, and Operations flashcards FAQ