90+ Cloud Security Interview Questions for DevOps (2027)
94 cloud security interview questions with answers for DevOps roles: IAM, OIDC in CI/CD, secrets, KMS, Kubernetes, supply chain and incident scenarios.
Cloud security interviews for DevOps roles test whether you can secure what you build and ship: IAM and least privilege, network controls, encryption and key management, secrets, logging and detection, pipeline and supply chain security, containers and Kubernetes, and incident response. This page has 94 questions with direct answers, grouped by experience level, plus scenario questions.
The questions here focus on the DevOps side of cloud security: pipelines, infrastructure as code, containers and day-to-day operations across AWS, Azure and Google Cloud. If you are interviewing for a dedicated security engineering role, pair this page with our cloud security engineer interview questions. For architect-level depth, read the advanced cloud security questions, and for long incident walkthroughs, the scenario-based cloud security questions. The CDN and edge side of the topic is covered in our Akamai interview questions.
Key takeaways
- Freshers are asked about the shared responsibility model, IAM basics, encryption at rest and in transit, security groups and logging.
- Engineers with 2 to 5 years are expected to secure CI/CD pipelines, remove long-lived keys with OIDC federation, scan infrastructure as code and container images, and lock down Kubernetes.
- Senior candidates are tested on multi-account guardrails, data perimeters, supply chain integrity (SBOMs, signing, SLSA) and detection engineering.
- Name the native service for each cloud when you answer, for example AWS KMS, Azure Key Vault and Google Cloud KMS, to show multi-cloud awareness.
- For scenario questions, answer in order: contain, investigate, fix, prevent.
How to use this page
| Your experience | Read first | Practise |
|---|---|---|
| Fresher or career switcher | Questions 1 to 35 | Create an IAM role, an encrypted bucket and a security group with the CLI; read a CloudTrail event |
| 2 to 5 years | Questions 36 to 70 | Set up GitHub Actions OIDC to a cloud role; add Trivy and Checkov to a pipeline |
| Senior | Questions 71 to 84 | Design an SCP and RCP set for a sandbox OU; sign an image with Cosign and verify it at admission |
| Everyone | Scenarios 85 to 94 | Talk through each one out loud in under three minutes |
Cloud security questions for freshers
1. What is the shared responsibility model?
The cloud provider secures the infrastructure that runs the cloud (data centres, hardware, hypervisor, managed service internals), and the customer secures what they put in it (data, identities, configuration, operating systems on VMs, application code). The split moves toward the provider as you go from IaaS to PaaS to SaaS, but data and access management always stay with the customer.
2. What is IAM?
Identity and access management controls who (users, groups, roles, service accounts) can do what (actions) on which resources, under which conditions. AWS IAM, Microsoft Entra ID with Azure RBAC, and Google Cloud IAM are the native services.
3. What is the principle of least privilege?
Give each identity only the permissions it needs for its task, for only as long as it needs them. In practice, start with narrow policies, use roles instead of broad user permissions, and remove access that logs show is unused.
4. What is the difference between authentication and authorization?
Authentication proves who you are (password plus MFA, a certificate, a federated token). Authorization decides what you are allowed to do once identified (IAM policies, RBAC roles).
5. Why should you not use the AWS root user or a global admin account for daily work?
These accounts have unrestricted access that policies cannot fully limit, so a compromise is total. Protect them with strong MFA, store credentials securely, use them only for the few tasks that require them, and alert on every sign-in.
6. What is MFA and which type is strongest?
Multifactor authentication requires a second factor beyond a password. Phishing-resistant methods such as FIDO2 security keys and passkeys are stronger than SMS codes or push notifications, which can be intercepted or approved by mistake.
7. What is the difference between an IAM user and an IAM role in AWS?
A user is a long-lived identity with permanent credentials. A role has no permanent credentials; it is assumed and returns temporary credentials from AWS STS. Prefer roles for workloads, pipelines and human access through federation.
8. What is encryption at rest?
Data is encrypted while stored on disks, in object storage or in databases, so stolen media or snapshots cannot be read without the keys. Most cloud storage services now encrypt at rest by default; your choice is usually which key to use.
9. What is encryption in transit?
Data is encrypted while moving across networks, normally with TLS. Enforce it by redirecting HTTP to HTTPS, requiring TLS on load balancers and denying unencrypted requests to storage.
10. What is a KMS?
A key management service creates, stores and controls the use of encryption keys, logs every use and supports rotation. Examples are AWS KMS, Azure Key Vault and Google Cloud KMS. Services call the KMS to encrypt data keys, so the master key never leaves it.
11. What is the difference between provider-managed and customer-managed keys?
Provider-managed keys are created and rotated by the service with no configuration. Customer-managed keys are keys you control: you set their access policy, rotation and deletion, and you can revoke access to data by disabling the key.
12. What is a security group?
A stateful virtual firewall attached to instances or network interfaces. You define allowed inbound and outbound traffic; return traffic for allowed connections is permitted automatically. Azure network security groups and Google Cloud firewall rules play the same role.
13. What is the difference between a security group and a network ACL in AWS?
Security groups are stateful, apply to network interfaces and support allow rules only. Network ACLs are stateless, apply to subnets, support allow and deny rules, and are evaluated in rule-number order.
14. Why use private subnets?
Resources in a private subnet have no direct route from the internet, which removes a large attack surface. Databases and application servers usually belong there, reached through load balancers or bastion alternatives.
15. How do private instances reach the internet for updates?
Through a NAT gateway that allows outbound connections only, or through VPC endpoints or Private Link to reach cloud services without the internet at all.
16. What is a VPC endpoint or Private Link?
A private connection from your virtual network to a cloud service or partner service that keeps traffic on the provider's network. It lets you block internet egress while still reaching storage, registries and APIs.
17. Why is opening SSH or RDP to 0.0.0.0/0 dangerous?
Internet-wide scanners find open management ports within minutes and try credential attacks and known vulnerabilities. Use session managers (AWS Systems Manager Session Manager, Azure Bastion, Google Cloud IAP) or restrict access to known addresses and require keys or MFA.
18. What is AWS CloudTrail and its equivalents?
CloudTrail records API calls made in an AWS account: who did what, when, from where. Azure Activity Log and Google Cloud Audit Logs do the same. These logs are the first source in any investigation.
19. Why send audit logs to a separate account or project?
An attacker who compromises a workload account may try to delete logs. Storing logs in a dedicated log archive account with restricted access and object lock or retention policies keeps them intact.
20. What is the difference between logging and monitoring?
Logging records events. Monitoring watches metrics and logs and alerts when something crosses a threshold or matches a pattern. You need both: logs for investigation, monitoring for timely detection.
21. What is a misconfiguration and why is it a common cloud risk?
A setting that weakens security, such as a public bucket, an open security group or a role with wildcard permissions. Cloud resources are created quickly through code and consoles, so one wrong setting can expose data immediately.
22. How do you prevent a storage bucket from being public?
Turn on account-level and bucket-level public access blocks (S3 Block Public Access, Azure storage "allow blob public access" disabled, Google Cloud public access prevention), avoid public ACLs, and scan for exceptions with posture tools.
23. What is CSPM?
Cloud security posture management continuously checks cloud configuration against best practice and compliance rules and reports misconfigurations. Examples include AWS Security Hub, Microsoft Defender for Cloud, Google Security Command Center and open-source tools like Prowler.
24. What is a secret in DevOps?
Any credential that grants access: passwords, API keys, tokens, private keys, database connection strings. Secrets must never be stored in Git, images or plain-text pipeline variables.
25. Where should secrets be stored?
In a secrets manager such as AWS Secrets Manager, Azure Key Vault, Google Secret Manager or HashiCorp Vault, with access controlled by IAM and every read logged. Applications fetch secrets at runtime.
26. What is secret scanning?
Automatic detection of credentials committed to code. GitHub secret scanning with push protection, GitLab secret detection and open-source tools like Gitleaks catch secrets before or after a push.
27. What is a WAF?
A web application firewall inspects HTTP requests and blocks common attacks such as SQL injection and cross-site scripting, and can rate-limit abusive clients. AWS WAF, Azure Web Application Firewall and Google Cloud Armor are native options.
28. What is DDoS protection in the cloud?
Services that absorb and filter volumetric and protocol attacks, such as AWS Shield, Azure DDoS Protection and Google Cloud Armor. Basic network-layer protection is included by default on major clouds; advanced tiers add application-layer protection and response support.
29. What is the CIA triad?
Confidentiality (only authorised people see data), integrity (data is not altered without authorisation) and availability (systems are usable when needed). Every control protects one or more of these.
30. What is a vulnerability scan?
An automated check for known vulnerabilities (CVEs) in operating systems, packages, container images or applications. Examples are Amazon Inspector, Defender for Cloud and Trivy.
31. What is patch management in the cloud?
Keeping operating systems and software updated. For VMs, use tools like AWS Systems Manager Patch Manager or Azure Update Manager. For containers, rebuild images from updated base images and redeploy rather than patching running containers.
32. What is a bastion host and what replaces it?
A hardened jump server for reaching private instances. Many teams replace it with identity-aware session services (Systems Manager Session Manager, Azure Bastion, Google Cloud IAP) that need no open inbound ports and log every session.
33. What is a service account?
A non-human identity used by applications and automation. Google Cloud and Kubernetes use the term directly; on AWS the equivalent is a role assumed by a workload, and on Azure a managed identity or service principal.
34. What is a managed identity in Azure?
An identity Azure creates and manages for a resource such as a VM or function, so code can get tokens for other Azure services without storing credentials. It can be system-assigned (tied to one resource) or user-assigned (shared).
35. What is data classification and why does it matter?
Labelling data by sensitivity (public, internal, confidential, restricted). It decides which encryption, access and retention controls apply, and tools such as Amazon Macie or Microsoft Purview help find sensitive data automatically.
Cloud security questions for 2 to 5 years of experience
36. How do you remove long-lived cloud keys from CI/CD?
Use OpenID Connect federation. The CI platform (GitHub Actions, GitLab CI, Azure DevOps) issues a short-lived signed token for each job, and the cloud trusts that token to grant temporary credentials for a specific role. No static keys are stored in the CI system.
37. How do you restrict which GitHub workflows can assume an AWS role through OIDC?
Add conditions to the role's trust policy on the token's aud claim (sts.amazonaws.com) and sub claim, which encodes the repository and the branch or environment, for example repo:my-org/my-repo:environment:production. Never trust a wildcard for the whole organisation.
"Condition": {
"StringEquals": {
"token.actions.githubusercontent.com:aud": "sts.amazonaws.com",
"token.actions.githubusercontent.com:sub": "repo:my-org/my-repo:environment:production"
}
}
38. What are the main risks in a CI/CD pipeline?
Stolen pipeline credentials, poisoned dependencies, malicious pull requests that run code with secrets, compromised third-party actions or plugins, tampered build artefacts and over-privileged deploy roles. OWASP's Top 10 CI/CD Security Risks is a good checklist.
39. How do you secure third-party GitHub Actions?
Pin actions to a full commit SHA instead of a tag, allow only approved actions through organisation policy, set the default GITHUB_TOKEN permissions to read-only, and grant extra permissions per job.
40. Why is pull_request_target risky in GitHub Actions?
It runs in the context of the base repository with access to secrets and a write token. If the workflow checks out and runs code from the pull request, an attacker's fork can execute code with those secrets. Use pull_request for untrusted code.
41. What is IaC scanning?
Static analysis of Terraform, CloudFormation, Bicep, Kubernetes YAML and Helm charts for misconfigurations before deployment. Checkov, Trivy (which absorbed tfsec) and KICS are common open-source scanners.
42. Where do you put security scans in a pipeline?
Secret scanning and IaC scanning on every pull request, SAST and dependency scanning on pull requests and main, image scanning after build, DAST against a staging deployment, and posture and runtime monitoring after release.
43. What is SAST, DAST and SCA?
SAST analyses source code for insecure patterns. DAST tests a running application from outside. SCA (software composition analysis) finds known vulnerabilities and licence issues in third-party dependencies.
44. How do you decide whether a vulnerability finding should block a build?
Block on critical and high findings that have a fix available and affect code that is actually used, and track the rest with an owner and a deadline. Blocking on every finding trains teams to ignore the scanner.
45. How do you secure container images?
Use minimal or distroless base images, pin base images by digest, run as a non-root user, scan images in CI and in the registry, rebuild regularly to pick up fixes, and sign images so deployments can verify them. See our container security tools guide.
46. What is image signing and how is it done?
Attaching a cryptographic signature to an image digest so consumers can prove who built it and that it was not changed. Sigstore Cosign is widely used, including keyless signing with OIDC identities. Admission controllers can verify signatures before pods run.
47. What are Kubernetes RBAC best practices?
Prefer namespaced Roles over ClusterRoles, avoid wildcard verbs and resources, do not bind cluster-admin to people or service accounts for routine work, give each workload its own service account, and review bindings regularly.
48. What are Pod Security Standards?
Three Kubernetes-defined profiles, Privileged, Baseline and Restricted, enforced by the built-in Pod Security Admission controller through namespace labels in enforce, audit or warn mode. Restricted blocks privileged containers, host namespaces and running as root.
49. Why use Kubernetes network policies?
By default all pods can talk to each other. NetworkPolicy resources restrict ingress and egress by labels, namespaces and ports, which limits lateral movement. They need a CNI plugin that enforces them, such as Calico or Cilium.
50. How should Kubernetes workloads access cloud APIs?
Through workload identity: EKS Pod Identity or IAM Roles for Service Accounts on AWS, Microsoft Entra Workload ID on AKS, and Workload Identity Federation for GKE on Google Cloud. Each maps a Kubernetes service account to a cloud identity with short-lived credentials, so no keys are stored in secrets.
51. Are Kubernetes Secrets encrypted?
Secrets are base64-encoded, not encrypted, by default in the API. Enable encryption at rest for etcd (managed services often support envelope encryption with a cloud KMS key), restrict RBAC on secrets, and consider an external secrets manager with the Secrets Store CSI driver or External Secrets Operator.
52. What is policy as code?
Writing security and compliance rules as code that tools evaluate automatically, for example OPA Rego, Kyverno policies, Kubernetes ValidatingAdmissionPolicy in CEL, or CloudFormation Guard rules. Policies are versioned and reviewed like application code. See our guide to tools to automate cloud security policies.
53. How do you rotate secrets without downtime?
Support two valid credentials during rotation: create the new one, update consumers to use it, verify, then revoke the old one. Managed rotation in AWS Secrets Manager and dynamic secrets in Vault automate this pattern.
54. What are dynamic secrets?
Credentials generated on demand for each client with a short lease, for example a database user created by HashiCorp Vault that expires after an hour. A leaked dynamic secret has a small window of use.
55. How do you enforce encryption on all new storage?
Turn on default encryption settings (for example EBS encryption by default), deny creation of unencrypted resources with organisation policies or IaC checks, and detect exceptions with posture rules.
56. What is envelope encryption?
Data is encrypted with a data key, and the data key is encrypted with a master key held in a KMS. Only the small data key travels to the KMS for decryption, which keeps large data local and the master key protected.
57. How do you detect threats in an AWS account?
Enable Amazon GuardDuty for threat detection from CloudTrail, VPC flow logs and DNS logs, aggregate findings in Security Hub, and route high-severity findings to alerting through EventBridge. Azure uses Defender for Cloud and Microsoft Sentinel; Google Cloud uses Security Command Center.
58. What are VPC flow logs used for?
They record metadata about accepted and rejected network connections (source, destination, port, bytes). They help detect scanning, unexpected egress and data exfiltration, and confirm whether a rule change took effect.
59. How do you limit the blast radius of a compromised deploy role?
Separate roles per environment and per pipeline, restrict actions to what the deployment needs, scope resources by tags or ARNs, add permission boundaries, and require environment approvals before the production role can be assumed.
60. What is a permissions boundary in AWS?
A managed policy that sets the maximum permissions an IAM user or role can have. It lets you delegate role creation to developers while guaranteeing the roles they create cannot exceed the boundary.
61. What is the difference between Azure RBAC and Microsoft Entra roles?
Azure RBAC roles control access to Azure resources (subscriptions, resource groups, VMs). Microsoft Entra roles control administration of the directory itself, such as managing users and applications.
62. What are Google Cloud IAM basic roles and why avoid them?
Owner, Editor and Viewer are broad legacy roles that apply across many services. Google recommends predefined or custom roles instead, because Editor in particular grants write access to most resources in a project.
63. How do you secure a container registry?
Make it private, restrict push to CI identities, enable scanning on push, use immutable tags, apply lifecycle policies to remove old images, and require signed images at deployment.
64. What is an SBOM?
A software bill of materials lists the components and versions inside an artefact. Formats include SPDX and CycloneDX, and tools like Syft or Trivy generate them. SBOMs let you answer "are we affected?" quickly when a new vulnerability is published.
65. How do you protect a public API on the cloud?
Put it behind an API gateway with authentication (OAuth 2.0 or JWT validation), rate limiting and request validation, add a WAF, use TLS everywhere, log every request, and keep backend services private.
66. What is the risk of instance metadata services and how do you reduce it?
The metadata endpoint can return temporary credentials for the instance role. A server-side request forgery bug could let an attacker read them. On AWS, require IMDSv2, which uses session tokens, and keep instance roles narrow.
67. How do you secure Terraform state?
Store it in a remote backend with encryption, versioning and locking, restrict access to the pipeline identity and a few administrators, and remember that state can contain secrets in plain text. Mark sensitive values and avoid writing secrets into resources where possible.
68. What is drift and why is it a security issue?
Drift is when live infrastructure differs from the code that should define it, often after a manual console change. It can silently reopen a port or widen a policy. Detect it with scheduled plans or GitOps controllers and remove console write access in production.
69. How do you give developers production access safely?
Default to no standing access. Use just-in-time elevation with approval and expiry (for example Microsoft Entra Privileged Identity Management or an access broker), session recording, read-only roles for routine debugging and break-glass accounts for emergencies.
70. What is zero trust in a cloud context?
Do not trust a request because of its network location. Verify identity, device and context on every request, give least privilege and assume breach. NIST SP 800-207 defines the architecture.
Senior cloud security questions
71. How would you design guardrails for a multi-account AWS organisation?
Group accounts into OUs by environment and function, apply SCPs to block dangerous actions (disabling CloudTrail, leaving the organisation, using unapproved regions), apply RCPs to limit access to resources from outside the organisation, centralise logs in a log archive account, delegate security tooling to a security account and enable detective controls everywhere.
72. What is the difference between SCPs and RCPs?
SCPs cap the permissions of IAM principals in member accounts. RCPs, introduced in November 2024, cap the permissions on resources in member accounts, including requests from principals outside the organisation. Neither grants access, and neither applies to the management account.
73. What is a data perimeter?
A set of preventive controls that ensure only trusted identities access trusted resources from expected networks. On AWS it combines SCPs, RCPs, resource policies and VPC endpoint policies with condition keys such as aws:PrincipalOrgID and aws:ResourceOrgID. On Google Cloud, VPC Service Controls play a similar role.
74. How do you secure the software supply chain end to end?
Protect source (branch protection, signed commits, reviews), protect the build (isolated ephemeral runners, pinned dependencies, no long-lived secrets), produce provenance and SBOMs, sign artefacts, verify signatures and provenance at deployment, and monitor dependencies for new vulnerabilities.
75. What is SLSA?
Supply-chain Levels for Software Artifacts is an OpenSSF framework that defines levels of build integrity, centred on generating trustworthy provenance about how an artefact was built. Higher build levels require hosted, isolated builds whose provenance cannot be forged by the build's own steps.
76. How do you approach key management for regulated workloads?
Use customer-managed keys with separation of duties between key administrators and key users, restrict key policies, enable rotation, log all key use and consider external key stores or HSM-backed keys where regulation requires that the provider cannot access key material.
77. How would you build a detection pipeline for cloud threats?
Collect audit, network, DNS and workload logs centrally; enable native threat detection; write detections mapped to MITRE ATT&CK for cloud; route alerts by severity to on-call; automate safe containment steps; and test detections regularly with controlled simulations.
78. What would you automate in incident response and what would you keep manual?
Automate low-risk, high-confidence actions such as disabling a leaked access key, quarantining an instance with a restrictive security group and snapshotting disks for forensics. Keep destructive or business-impacting decisions, such as deleting resources or shutting down a production service, behind a human decision.
79. How do you secure a multi-tenant Kubernetes platform?
Separate tenants by namespace with RBAC, quotas, network policies and Pod Security Standards; use admission policies for guardrails; give each tenant its own workload identity; and use separate clusters or node pools when tenants need stronger isolation.
80. How do you handle security for serverless workloads?
Give each function its own narrow role, validate input from every trigger, keep secrets in a secrets manager, limit concurrency and timeouts to reduce abuse cost, scan dependencies, and log invocations with enough context to investigate.
81. How do you measure cloud security posture over time?
Track a small set of metrics: critical misconfigurations open and mean time to fix, percentage of identities without MFA, number of long-lived access keys, critical vulnerabilities past SLA, and coverage of logging and detection across accounts.
82. How do you balance developer speed with security controls?
Provide secure defaults through templates and platform modules, run checks automatically in pipelines with fast feedback, start new policies in audit mode, and reserve hard blocks for high-risk issues. Make the secure way the easy way.
83. How do you prepare for a cloud provider region outage from a security angle?
Replicate keys, secrets and logging to the recovery region, make sure IAM and guardrails apply there too, test failover with security controls enabled, and keep break-glass access that does not depend on the failed region.
84. How do compliance frameworks map to DevOps controls?
Frameworks such as ISO/IEC 27001, SOC 2 and PCI DSS describe outcomes, such as access control, change management and logging. DevOps teams meet them with code review, pipeline approvals, IaC with policy checks, central logging and automated evidence. See our DevOps compliance tools guide.
Scenario questions
85. A developer pushed an AWS access key to a public GitHub repository. What do you do?
Deactivate the key immediately, then delete it after confirming no production dependency remains. Review CloudTrail for every call made with that key since the push, look for new users, roles, keys or resources created by it, and remove anything unexpected. Rewrite history if needed, but treat the key as compromised regardless. Prevent repeats with push protection, OIDC federation for pipelines and an SCP or alert on access key creation.
86. Security tooling reports an S3 bucket with customer data is public. What are your steps?
Block public access on the bucket at once, check server access logs and CloudTrail data events for reads during the exposure window, determine what data was exposed and involve legal and privacy teams if personal data was read. Then turn on account-level Block Public Access and add an IaC check so the setting cannot return.
87. GuardDuty reports an EC2 instance communicating with a known cryptomining domain. What do you do?
Isolate the instance with a quarantine security group, snapshot its volumes for forensics, revoke the instance role's active sessions, and check CloudTrail for actions taken with the role. Find the entry point (vulnerable software, exposed port, stolen key), rebuild from a clean image, patch the cause and add detection for similar behaviour.
88. A pipeline must deploy to production in three cloud accounts. How do you design its access?
Use OIDC from the CI platform to a deploy role in each account, with trust conditions limited to the repository and the production environment. Scope each role to the services the deployment touches, require environment approval before the production job runs, and log role assumptions centrally.
89. Your Kubernetes cluster is running pods from an image nobody recognises. How do you respond?
Find which workload created the pods and which identity applied it using Kubernetes audit logs. Scale it down or delete it, capture the image and pod details for analysis, rotate any secrets the pod could read, and review RBAC for the identity that created it. Prevent repeats with admission policies that allow only signed images from approved registries.
90. A Terraform plan shows a security group rule opening port 22 to the world. The author says it is temporary. What do you do?
Do not approve it. Offer a session manager or IAP-based access instead, which needs no inbound port. If an exception is unavoidable, restrict the source to known addresses, set an expiry and track it. Add a policy check so this pattern fails the plan automatically.
91. Audit finds dozens of IAM roles with AdministratorAccess. How do you fix this without breaking teams?
Use IAM Access Analyzer or last-accessed data to see which services each role actually uses, generate narrower policies from that activity, roll them out in stages with monitoring for access-denied errors, and set a permissions boundary for new roles. Remove the admin policy once the replacement is stable.
92. You discover a third-party GitHub Action used in your pipelines was compromised. What now?
Identify every workflow that ran the affected versions and when, assume any secrets available to those jobs were exposed and rotate them, check logs for unexpected outbound connections or artefact changes, and rebuild any artefacts produced during the window. Then pin all actions to full commit SHAs and restrict allowed actions at the organisation level.
93. A database password is hard-coded in a container image already in production. How do you handle it?
Rotate the database password first, because anyone with registry access may have pulled the image. Move the secret to a secrets manager and inject it at runtime, rebuild and redeploy the image, delete the old image tags from the registry and add secret scanning to the image build.
94. Logging costs are high and a manager wants to turn off CloudTrail data events and flow logs. How do you respond?
Explain which investigations depend on those logs and propose cheaper options rather than switching them off: log only sensitive buckets and critical subnets, sample or filter flow logs, move older logs to cheaper storage tiers and shorten retention where policy allows. Keep management events and logs for regulated data.
How to prepare
- Build one small project end to end: a GitHub repository that deploys to a cloud account through OIDC, with secret scanning, Checkov, Trivy and an environment approval.
- Read real cloud audit log events and practise explaining what each field means.
- Learn the native service names for IAM, keys, secrets, logs and threat detection on all three major clouds.
- Practise scenarios aloud with the contain, investigate, fix, prevent structure.
- Review the broader 200+ DevOps interview questions for the Linux, Git and Kubernetes rounds around security.
Our DevSecOps training covers pipeline security, cloud IAM, container and Kubernetes security with labs, in the Kharadi, Pune classroom or live online.
Frequently asked questions
Which cloud security topics are asked most in DevOps interviews?
IAM and least privilege, secrets management, CI/CD pipeline security, encryption with KMS, network controls, container and Kubernetes security, and logging. Expect at least one incident scenario.
Do I need a security certification to get these jobs?
It is not required by every employer. Hands-on experience with pipelines, IAM and Kubernetes security usually matters more in DevOps interviews, and certifications can help you get shortlisted. Check each certifying body's site for current exams.
How do I answer a cloud security scenario question?
Use four steps: contain the issue, investigate scope with logs, fix the root cause, and prevent it with an automated control. Name the specific services you would use.
What is OIDC federation in CI/CD?
The CI platform issues a short-lived token per job and the cloud exchanges it for temporary credentials for a specific role. It removes stored cloud keys from pipelines.
Should I learn AWS, Azure or Google Cloud security first?
Learn the one your target employers use most, then map the concepts to the others. IAM, keys, logging and network controls exist on all three under different names.
How is this page different from the advanced cloud security questions?
This page covers the full range from fresher to senior with a DevOps and pipeline focus. The advanced page goes deeper into architecture and design questions for experienced candidates.
Official references
- AWS Shared Responsibility Model
- GitHub Docs: Security hardening with OpenID Connect
- Kubernetes: Pod Security Standards
- OWASP Top 10 CI/CD Security Risks
- SLSA framework
- NIST SP 800-207: Zero Trust Architecture
- AWS Organizations: Resource control policies
- Microsoft Learn: Shared responsibility in the cloud
- Google Cloud IAM overview
What's Your Reaction?
Like
0
Dislike
0
Love
0
Funny
0
Angry
0
Sad
0
Wow
0