70+ Argo CD Interview Questions and Answers (2027)

75 Argo CD interview questions with answers for freshers, 2-5 years and seniors: sync, prune, self-heal, RBAC, ApplicationSets, hooks and 3.0 changes.

Sep 17, 2025 - 11:24
Nov 6, 2026 - 01:36
 0  246
70+ Argo CD Interview Questions and Answers (2027)

Argo CD interview questions check whether you understand GitOps delivery on Kubernetes: how Argo CD compares Git with the live cluster, how sync, prune and self-heal work, how Projects, RBAC and SSO control access, and how ApplicationSets, sync waves and hooks handle many apps and ordered rollouts. This page has 75 numbered questions with direct answers, grouped by experience level.

Answers were checked against the official Argo CD documentation in November 2026, including the changes that arrived with Argo CD 3.0. For incident-style questions only, see our separate scenario-based Argo CD interview questions; this page focuses on concepts, configuration and design.

Key takeaways

  • Argo CD is a pull-based GitOps controller. It runs inside Kubernetes, reads desired state from Git (or Helm and OCI registries) and reconciles the cluster toward it.
  • Freshers should explain the Application resource, sync status versus health status, and the difference between manual and automated sync.
  • Mid-level candidates are asked about Projects, RBAC in argocd-rbac-cm, sync waves, hooks, ignoreDifferences and ApplicationSet generators.
  • Senior candidates discuss multi-cluster topology, controller sharding, secrets handling and the Argo CD 3.0 upgrade (annotation tracking by default, logs RBAC enforced, fine-grained RBAC).
  • Argo CD is a CNCF graduated project. Argo Rollouts, Argo Workflows and Argo Events are separate projects under the same umbrella.

How to use this list

Your experienceQuestionsHands-on practice
Fresher or career switcher1 to 25Install Argo CD on kind or minikube and deploy the guestbook example
2 to 5 years26 to 50Build an AppProject with RBAC, add sync waves and a PreSync hook
Senior or platform engineer51 to 67Manage two clusters with one ApplicationSet and test controller sharding
Everyone68 to 75Talk through each design scenario out loud in under three minutes

If GitOps itself is new to you, read what GitOps is and how it helps infrastructure automation before you start.

Argo CD interview questions for freshers

1. What is Argo CD?

Argo CD is an open source, declarative continuous delivery tool for Kubernetes that follows the GitOps model. It watches Git repositories that hold Kubernetes manifests, compares them with what is running in the cluster, and reports or fixes any difference.

2. What is GitOps, and how does Argo CD implement it?

GitOps means the desired state of a system is declared, versioned in Git, pulled automatically by an agent and continuously reconciled. Argo CD is that agent: its application controller pulls manifests and keeps the cluster matching them. The OpenGitOps project publishes these four principles.

3. What is the difference between push-based and pull-based deployment?

In push-based delivery a CI server runs kubectl apply or helm upgrade against the cluster, so it needs cluster credentials. In pull-based delivery an agent inside the cluster pulls from Git. Argo CD is pull-based, which keeps cluster credentials out of the CI system.

4. What are the main components of Argo CD?

The API server (argocd-server) serves the web UI, CLI and gRPC/REST API. The repo server (argocd-repo-server) clones repositories and renders manifests. The application controller compares live and desired state and runs syncs. Redis is used as a cache, Dex is optional for SSO, and the ApplicationSet and notifications controllers ship with the standard install.

5. What is an Application in Argo CD?

An Application is a Kubernetes custom resource (argoproj.io/v1alpha1, kind Application) that links a source (repo URL, revision, path or chart) to a destination (cluster and namespace). It also holds the project name and sync policy.

apiVersion: argoproj.io/v1alpha1
kind: Application
metadata:
  name: guestbook
  namespace: argocd
spec:
  project: default
  source:
    repoURL: https://github.com/argoproj/argocd-example-apps.git
    targetRevision: HEAD
    path: guestbook
  destination:
    server: https://kubernetes.default.svc
    namespace: guestbook

6. What does "sync" mean?

Sync is the operation that applies the desired manifests from Git to the cluster so the live state matches the target state. A sync can be started manually from the UI or CLI, or automatically by a sync policy.

7. What is the difference between sync status and health status?

Sync status says whether live resources match Git: Synced, OutOfSync or Unknown. Health status says whether the resources are working: Healthy, Progressing, Degraded, Suspended, Missing or Unknown. An app can be Synced yet Degraded, for example when a Deployment matches Git but its pods crash.

8. What does OutOfSync mean?

At least one resource in the cluster differs from the rendered manifests, or a resource in Git does not exist in the cluster yet. Common causes are a new commit, a manual kubectl edit, or a field changed by another controller.

9. How do you install Argo CD?

Create the argocd namespace and apply the official install manifest, or use the community Helm chart.

kubectl create namespace argocd
kubectl apply -n argocd -f https://raw.githubusercontent.com/argoproj/argo-cd/stable/manifests/install.yaml

For production, the docs recommend the HA manifests, which run multiple replicas and Redis HA.

10. How do you get the initial admin password?

Argo CD stores it in the argocd-initial-admin-secret secret. Read it with argocd admin initial-password -n argocd, log in, change the password, then delete the secret. Most teams disable the local admin user once SSO works.

11. How do you reach the Argo CD UI on a test cluster?

Port-forward the server service with kubectl port-forward svc/argocd-server -n argocd 8080:443 and open https://localhost:8080. In real environments you expose it through an Ingress or a LoadBalancer with TLS.

12. Which CLI commands should a beginner know?

argocd login, argocd app create, argocd app list, argocd app get, argocd app sync, argocd app diff, argocd app history, argocd app rollback, argocd repo add and argocd cluster add.

13. Which manifest formats does Argo CD support?

Plain YAML or JSON directories, Kustomize, Helm charts and Jsonnet. Anything else can be rendered through a Config Management Plugin that runs as a sidecar on the repo server.

14. How does Argo CD use Helm?

Argo CD renders the chart with helm template and applies the output itself. Releases do not show up in helm list, because Argo CD, not Helm, owns the lifecycle. Values come from valueFiles, values, valuesObject or parameters in the Application. For chart design questions, see our Helm interview questions.

15. What does targetRevision control?

It selects what to deploy from the source: a branch, a tag, a commit SHA or HEAD for Git, or a chart version for Helm repositories. Pinning a tag or SHA gives repeatable deployments; tracking a branch gives continuous delivery.

16. What is automated sync?

With syncPolicy.automated set, Argo CD syncs whenever it detects that Git has changed. Without it, the app shows OutOfSync and waits for someone to sync it.

17. What do prune and self-heal do?

prune: true lets automated sync delete resources that were removed from Git. selfHeal: true makes Argo CD revert manual changes in the cluster even when Git has not changed. Both are off by default.

syncPolicy:
  automated:
    prune: true
    selfHeal: true
  syncOptions:
    - CreateNamespace=true

18. How often does Argo CD check Git?

By default it polls repositories about every three minutes (the timeout.reconciliation setting, 180 seconds plus jitter). Configure a Git webhook to the /api/webhook endpoint so changes are picked up almost at once.

19. What is the difference between refresh and hard refresh?

A normal refresh compares Git with the live state using cached manifests where possible. A hard refresh clears the manifest cache and re-renders from the repository, which helps when a Helm dependency or plugin output changed without a new commit.

20. How do you roll back an application?

Use argocd app history to find a previous deployment ID and argocd app rollback <app> <id>. Rollback is blocked while automated sync is enabled, because Argo CD would sync back to Git. The cleaner GitOps rollback is a git revert.

21. What is an AppProject?

An AppProject groups Applications and limits what they can do: which source repos they may use, which clusters and namespaces they may deploy to, which resource kinds are allowed, and which roles can act on them. Every install has a default project with no restrictions.

22. How do you add a private Git repository?

Run argocd repo add with an SSH key, HTTPS token or GitHub App credentials, or create a Kubernetes Secret labelled argocd.argoproj.io/secret-type: repository. Credential templates (repo-creds) let one secret cover every repo under a URL prefix.

23. How does Argo CD deploy to another cluster?

Register the cluster with argocd cluster add <kube-context>. This creates a service account in the target cluster and stores its credentials as a cluster secret in the argocd namespace. Applications then set destination.server or destination.name to that cluster.

24. What is the App of Apps pattern?

One parent Application points at a Git directory that contains other Application manifests. Syncing the parent creates the children, which makes it easy to bootstrap a whole cluster from a single resource.

25. How is Argo CD different from Jenkins or GitHub Actions?

Jenkins and GitHub Actions are CI tools that build, test and push images. Argo CD only handles delivery to Kubernetes. A common split is CI building the image and updating a tag in a Git config repo, then Argo CD deploying that commit.

Argo CD interview questions for 2 to 5 years of experience

26. What are sync waves?

Sync waves order resources inside one sync. You set the annotation argocd.argoproj.io/sync-wave: "<number>"; lower numbers apply first and the default is 0. Argo CD waits for each wave to be healthy before starting the next.

27. What are resource hooks?

Hooks are resources, usually Jobs, that run at a set point in the sync. You mark them with argocd.argoproj.io/hook. Supported values are PreSync, Sync, PostSync, SyncFail, PostDelete and Skip.

28. Give a typical use of a PreSync hook.

Running database schema migrations before the new application version starts. If the migration Job fails, the sync stops and the old version keeps running.

apiVersion: batch/v1
kind: Job
metadata:
  name: db-migrate
  annotations:
    argocd.argoproj.io/hook: PreSync
    argocd.argoproj.io/hook-delete-policy: BeforeHookCreation
spec:
  template:
    spec:
      restartPolicy: Never
      containers:
        - name: migrate
          image: example/app-migrations:1.4.0
          command: ["./migrate", "up"]

29. What are hook delete policies?

The annotation argocd.argoproj.io/hook-delete-policy decides when hook resources are removed: HookSucceeded, HookFailed or BeforeHookCreation. Without a policy, hook resources stay until the next run replaces them.

30. Name some useful sync options.

Sync optionEffect
CreateNamespace=trueCreates the destination namespace if it is missing
PruneLast=trueDeletes pruned resources after all other waves finish
ApplyOutOfSyncOnly=trueApplies only changed resources, which helps large apps
ServerSideApply=trueUses Kubernetes server-side apply, useful for large CRDs
Replace=trueUses replace instead of apply
Prune=false (annotation)Protects a single resource from pruning
RespectIgnoreDifferences=trueKeeps ignored fields untouched during sync

31. What is ignoreDifferences for?

It tells Argo CD to ignore fields that legitimately differ from Git, such as spec.replicas managed by a HorizontalPodAutoscaler or a CA bundle injected into a webhook. You can match fields with jsonPointers, jqPathExpressions or managedFieldsManagers.

spec:
  ignoreDifferences:
    - group: apps
      kind: Deployment
      jsonPointers:
        - /spec/replicas

32. How does Argo CD RBAC work?

Policies live in the argocd-rbac-cm ConfigMap in Casbin CSV format. Each line grants a role or group an action on a resource, for example p, role:dev, applications, sync, team-a/*, allow. policy.default sets the fallback role, often role:readonly.

33. What changed for RBAC in Argo CD 3.0?

Three things are commonly asked. Log access now always requires an explicit logs, get permission. The update and delete permissions on an Application no longer cover its child resources unless you grant update/* or delete/*. With Dex, RBAC subjects match the federated_claims.user_id claim instead of sub.

34. How do you set up SSO?

Either use the bundled Dex, which connects to GitHub, LDAP, SAML and other providers, or configure an OIDC provider directly with oidc.config in argocd-cm. Map identity provider groups to Argo CD roles with g, lines in the RBAC policy.

35. What are project roles and JWT tokens?

An AppProject can define roles with their own policies. You can issue tokens for those roles with argocd proj role create-token, which gives CI jobs narrowly scoped access, such as syncing only that project's apps.

36. What are sync windows?

Sync windows, set on an AppProject, allow or deny syncs during a cron-based schedule and duration. A deny window over a weekend freeze stops automated syncs; you can choose whether manual syncs are still permitted.

37. What is an ApplicationSet?

An ApplicationSet is a custom resource that generates many Applications from one template. Generators supply parameters such as cluster names or Git directories, and the controller creates, updates and deletes the Applications to match.

38. Which ApplicationSet generators exist?

List, Cluster, Git (directory and file), Matrix, Merge, SCM Provider, Pull Request, Cluster Decision Resource and Plugin. Matrix and Merge combine other generators.

39. When would you choose ApplicationSet over App of Apps?

ApplicationSet fits when the Applications follow one pattern, such as the same app on every cluster or one app per folder. App of Apps fits when each child is hand-written and different. Many platforms bootstrap with App of Apps and use ApplicationSets underneath.

40. How does the Pull Request generator help?

It lists open pull requests from GitHub, GitLab, Bitbucket, Gitea or Azure DevOps and creates a preview Application per PR. When the PR closes, the Application and its environment are removed.

41. What is resource tracking, and what changed in 3.0?

Argo CD needs to know which live resources belong to which Application. Older versions used the app.kubernetes.io/instance label by default. Since 3.0 the default is the argocd.argoproj.io/tracking-id annotation, which avoids clashes with Helm and other tools that set the same label. You can switch back with application.resourceTrackingMethod.

42. How do you write a custom health check?

Add a Lua script under resource.customizations.health.<group>_<kind> in argocd-cm. The script reads the object and returns a status and message. Custom health checks matter for CRDs that Argo CD does not understand out of the box.

43. What is a Config Management Plugin?

A plugin lets the repo server render manifests with a tool Argo CD does not support natively. Plugins run as sidecar containers on the repo server with a plugin.yaml config. The older method of declaring plugins in argocd-cm was removed in version 2.8.

44. What are multiple sources?

An Application can use spec.sources to combine several sources, for example a Helm chart from a public repository with a values file from your own Git repo using a ref source.

45. How does Argo CD deal with secrets?

It does not encrypt secrets itself. Common patterns are Sealed Secrets, the External Secrets Operator pulling from Vault or a cloud secret manager, or SOPS through a plugin. The docs prefer operators that create secrets in the cluster over injecting secrets during rendering.

46. What is the resources finalizer?

Adding resources-finalizer.argocd.argoproj.io to an Application makes deleting the Application also delete the resources it deployed (cascading delete). Without it, deleting the Application leaves the workloads running.

47. What does selective sync do?

It syncs only chosen resources, for example argocd app sync myapp --resource apps:Deployment:web. It is handy for urgent fixes, but hooks do not run and the sync is not recorded in history.

48. What are Argo CD Notifications?

A controller bundled with Argo CD that sends messages on events such as sync succeeded, sync failed or health degraded. Triggers and templates are set in argocd-notifications-cm, and Applications subscribe with annotations. Slack, email, Teams, webhooks and many other services are supported.

49. How do Argo CD and Argo Rollouts relate?

Argo Rollouts is a separate controller that adds canary and blue-green strategies through a Rollout resource. Argo CD deploys the Rollout like any other manifest and understands its health, including a paused state.

50. What is Argo CD Image Updater?

A separate Argo project that watches container registries for new image tags and updates Applications, either through parameter overrides or by committing back to Git. It is useful when you do not want CI to write to the config repo.

Argo CD interview questions for senior engineers

51. Hub-and-spoke or one Argo CD per cluster?

A central instance managing many clusters gives one UI and one place for policy, but it is a single point of failure and needs network access to every API server. One instance per cluster isolates failures and scales simply, but multiplies upgrades and config. Many organisations use one instance per environment or region as a middle ground.

52. How do you scale the application controller?

Run the controller StatefulSet with more replicas and set ARGOCD_CONTROLLER_REPLICAS. Clusters are sharded across replicas using the legacy, round-robin or consistent-hashing algorithm. Also tune --status-processors and --operation-processors for many Applications.

53. How do you scale the repo server?

Add replicas, raise --parallelismlimit carefully and give it enough CPU and memory, since Helm and Kustomize rendering is CPU-heavy. Monorepos benefit from the manifest-generate-paths annotation, which limits refreshes to apps whose paths changed.

54. What is "Applications in any namespace"?

By default Application resources must live in the Argo CD namespace. Setting application.namespaces lets teams create Applications in their own namespaces, with AppProject sourceNamespaces controlling which namespaces may use which project. It suits self-service multi-tenancy.

55. How do you design multi-tenancy with Argo CD?

Give each team an AppProject that restricts source repos, destination namespaces and cluster-scoped kinds, map SSO groups to project roles, and deny cluster-scoped resources such as ClusterRoles to tenants. Keep platform add-ons in a separate admin project.

56. How should you structure Git repositories?

Most teams keep application source and deployment config in separate repos, so a config change does not trigger CI and image bumps do not clutter code history. Environments are usually folders (Kustomize overlays or values files) rather than long-lived branches, because folders are easier to diff and promote.

57. How do you promote a release across environments?

Change the image tag or chart version in the dev folder, then open a pull request that applies the same change to staging and then production. Each environment's Application tracks its own folder, so promotion is a reviewed Git change.

58. How do you let Argo CD manage itself?

Install Argo CD once, then create an Application that points at the Git folder holding Argo CD's own manifests or Helm values. Later upgrades and config changes go through Git. Keep a documented manual recovery path in case a bad change breaks the server.

59. How do you secure an Argo CD installation?

Disable the admin account after SSO works, use least-privilege RBAC with a read-only default, restrict AppProjects, keep the API server behind TLS and network controls, store repo and cluster credentials as secrets with tight access, and stay on a supported version for security fixes. Argo CD publishes advisories on its GitHub security page.

60. Why is the default project risky?

It allows any repo, any destination and any resource kind. If a tenant can create Applications in it, they can deploy cluster-scoped resources. Many teams restrict or empty the default project and require explicit projects.

61. How do you monitor Argo CD?

Each component exposes Prometheus metrics. Watch argocd_app_info labels for sync and health status, reconciliation duration, Git request counts and repo server render times. Argo CD 3.0 removed the older argocd_app_sync_status and argocd_app_health_status metrics, so dashboards built on them need updating.

62. What is server-side diff?

Instead of computing diffs locally, Argo CD can ask the Kubernetes API for a dry-run server-side apply and compare the result. It handles defaulting and mutating webhooks more accurately. You enable it per Application with an annotation or globally in argocd-cmd-params-cm.

63. How do you handle CRDs and their custom resources in one app?

Put the CRDs in an earlier sync wave than the custom resources, or in a separate Application that syncs first. Large CRDs often need ServerSideApply=true because client-side apply stores the whole object in an annotation with a size limit.

64. What would you check before upgrading from 2.x to 3.x?

Read the official upgrade notes. Plan for annotation-based tracking, the RBAC changes for logs and child resources, repositories declared in argocd-cm no longer being supported (move them to secrets), removed metrics and nested ApplicationSet selectors always applying. Upgrade one minor version at a time in staging first.

65. How do you deal with drift caused by other controllers?

Find which manager owns the field, then either remove that field from Git, add ignoreDifferences (often with managedFieldsManagers) or enable server-side diff. Avoid turning off self-heal across the board just to hide one noisy field.

66. How do you test Argo CD changes before merging?

Render manifests in CI with helm template or kustomize build, validate them with schema and policy tools, and run argocd app diff with a read-only token to show what would change. Pull Request generator previews add a live check.

67. What is the disaster recovery plan for Argo CD?

Git already holds the desired state, so the main extra items are Argo CD config, projects, repo and cluster secrets. argocd admin export and argocd admin import back up and restore these. Rebuilding the cluster, reinstalling Argo CD and re-applying the root Application should bring workloads back.

Scenario-based Argo CD design questions

These test judgement. For troubleshooting drills with symptoms and fixes, use the scenario-based Argo CD guide.

68. You must deploy the same add-ons to 40 clusters. What do you build?

An ApplicationSet with a Cluster generator that selects clusters by label, plus a Matrix or Git generator for the add-on list. New clusters get add-ons as soon as they are registered with the right label.

69. An HPA keeps scaling and the app is always OutOfSync. What do you do?

Remove replicas from the Deployment in Git or add an ignoreDifferences entry for /spec/replicas, together with RespectIgnoreDifferences=true so sync does not reset it.

70. A migration must run before the new pods start, and the app must not upgrade if it fails. How?

Use a PreSync hook Job for the migration with BeforeHookCreation as delete policy. A failed hook fails the sync, so the Deployment is never updated. Add a SyncFail hook to send an alert or clean up.

71. Developers keep fixing production with kubectl edit. How do you stop it?

Turn on selfHeal so manual changes are reverted, remove direct write access through Kubernetes RBAC, and give developers a fast pull request path. Notifications on OutOfSync events show who is drifting.

72. Production deploys must stop during a month-end freeze. How?

Add a deny sync window to the production AppProject for the freeze dates. Decide whether manualSync stays allowed for emergency fixes, and document who may use it.

73. Your config repo has 2,000 apps and Argo CD is slow. Where do you look?

Check repo server CPU and render time, Git request rates and controller queue depth. Add webhooks instead of relying on polling, use manifest-generate-paths annotations, shard the controller and scale the repo server.

74. A team wants preview environments for every pull request. What do you propose?

An ApplicationSet with a Pull Request generator that templates the namespace and image tag from the PR number and head SHA. Set resource quotas on preview namespaces and rely on the generator to delete them when PRs close.

75. Security asks how a compromised CI system could affect production. What is your answer?

With Argo CD, CI only writes to Git, so an attacker would have to get a commit merged past branch protection and review. Strengthen this with protected branches, required reviews, signed commits where Argo CD verifies GnuPG signatures for the project, and scoped project tokens.

Argo CD study plan

  1. Install Argo CD on a local cluster and deploy the example guestbook app.
  2. Convert it to Helm and Kustomize, and compare how Argo CD renders each.
  3. Add automated sync with prune and self-heal, then edit a resource by hand and watch it revert.
  4. Create an AppProject with SSO roles, a sync window and a restricted destination.
  5. Write an ApplicationSet for two clusters and a PR preview generator.
  6. Read the 2.14 to 3.0 upgrade notes so you can answer upgrade questions.

Argo CD sits on top of Kubernetes, so interviewers also test cluster basics; our Kubernetes architecture components guide and Kubernetes scenario questions cover that ground. For the wider pipeline picture, see CI/CD pipeline examples, and for every other topic start at the DevOps interview questions hub.

If you want guided labs, the CKA training builds the Kubernetes base, and the DevOps training covers GitOps pipelines. Both run in our Kharadi, Pune classroom and live online.

FAQ

Is Argo CD hard to learn for an interview?

Not if you know Kubernetes basics. Most of Argo CD is the Application resource, sync policies and Projects. A week of hands-on practice on a local cluster covers the fresher and mid-level questions.

What is the difference between Argo CD and Flux?

Both are CNCF graduated GitOps tools. Argo CD ships a web UI, Projects and SSO-backed RBAC in one install. Flux is a set of smaller controllers driven mostly through CRDs and the CLI, with an optional UI from third parties.

Does Argo CD replace CI tools?

No. Argo CD only delivers manifests to Kubernetes. You still need CI such as Jenkins, GitLab CI or GitHub Actions to build, test and publish images.

Which Argo CD version should I study?

Study the current 3.x release line and know the main 3.0 changes: annotation-based tracking, enforced logs RBAC and fine-grained RBAC for child resources.

What is the most common Argo CD interview question?

Explaining the difference between sync status and health status, followed by how prune and self-heal work in automated sync.

Are scenario questions asked for Argo CD roles?

Yes, especially for 3+ years of experience. Expect questions on drift, failed hooks, slow syncs and multi-cluster design, which our scenario guide covers in depth.

Do I need to know Helm for Argo CD interviews?

Yes. Many teams deploy Helm charts through Argo CD, so you should know how values are passed and why releases do not appear in helm list.

References

What's Your Reaction?

Like Like 0
Dislike Dislike 0
Love Love 0
Funny Funny 0
Angry Angry 0
Sad Sad 0
Wow Wow 0
Mridul I am a passionate technology enthusiast with a strong focus on DevOps, Cloud Computing, and Cybersecurity. Through my blogs at DevOps Training Institute, I aim to simplify complex concepts and share practical insights for learners and professionals. My goal is to empower readers with knowledge, hands-on tips, and industry best practices to stay ahead in the ever-evolving world of DevOps.