Cloud concept · Updated

Credential Management - Credential management is the controlled creation, storage, use, renewal, and revocation of the authentication material that people and workloads use to access systems.

In short

Credentials include passwords, access keys, session tokens, and other material used to prove an identity. Managing their lifecycle reduces credential theft, limits the impact of compromise, and keeps applications working when credentials change or expire.

Cloud Concepts Hub →

6 key aspects of Credential Management

Effective credential management minimizes persistent secrets, protects authentication paths, and makes credential replacement and revocation reliable.

01
Credential lifecycle control

Track credential ownership, purpose, consumers, and retirement requirements. Remove unused credentials and revoke exposed credentials rather than merely deleting their visible copies.

02
Temporary credentials by default

Use workload roles and federated sessions instead of distributing long-lived access keys. Ensure applications refresh expiring credentials through supported SDK credential providers.

03
Protected storage and retrieval

Keep secrets out of source code, images, and logs. Use encrypted storage with narrowly scoped retrieval permissions, and protect secrets after they enter application memory.

04
Password and MFA controls

Apply password requirements in the identity system that authenticates the user, and enable MFA where supported. Treat console passwords and programmatic credentials as separate authentication methods with separate controls.

05
Root and emergency access protection

Reserve root access for tasks that require it, avoid root access keys, and protect root sign-in with MFA and secured recovery channels. Monitor its use and maintain a controlled emergency-access procedure.

06
Safe rotation and auditability

Coordinate credential replacement with dependent applications to avoid outages. Record relevant authentication and secret-access events, and investigate unexpected use or failed retrievals.

Cloud services for Credential Management

IAM and IAM Identity Center establish how people and workloads obtain access, while AWS STS supplies temporary credentials for role-based sessions. Applications can use those credentials to retrieve application secrets from Secrets Manager or encrypted configuration from Systems Manager Parameter Store without embedding AWS access keys. IAM permissions constrain which identities can retrieve each value; AWS KMS provides the underlying encryption controls for these stores. CloudTrail supplies an audit trail for relevant management and retrieval activity. Together, these services separate identity access from application secrets, reduce persistent credentials, and support controlled rotation and incident investigation.

AWS Secrets ManagerAWS Secrets Manager

Secrets Manager stores secrets such as database credentials and API tokens, encrypts them, and controls retrieval through permissions. It supports automatic rotation through supported integrations or configured rotation workflows; rotation must also update the corresponding target system.

AWS Systems ManagerAWS Systems Manager

Systems Manager Parameter Store provides centralized configuration storage, including SecureString parameters encrypted with AWS KMS. It can hold sensitive configuration values but does not offer the same built-in secret rotation capabilities as Secrets Manager.

AWS CloudTrailAWS CloudTrail

CloudTrail records AWS account activity, including relevant credential-management operations, secret-retrieval API calls, and root activity. These events support investigations and can feed monitoring rules for unexpected credential or privileged-account use.

Credential Management exam questions

Real practice exam questions on Credential Management, with the correct answer and its explanation. Answer them with every option in the exam simulator.

MLA-C02 (beta)

An ML team will run a 72-hour unattended evaluation of an Amazon Bedrock foundation model by using a third-party testing appliance. The appliance supports Bedrock inference through HTTPS bearer-token authentication but cannot sign AWS requests. It loads one credential from an encrypted secret store when the evaluation starts and cannot refresh or replace that credential during the run. Company policy permits a dedicated, narrowly scoped long-lived credential for this integration, provided that it is revoked after the evaluation. All required model access is configured.

Which credential configuration meets these requirements?

Show answerCORRECT ANSWER

Create a long-term Amazon Bedrock API key associated with a dedicated IAM user that has only the required model invocation permissions. Set the key to remain valid throughout the evaluation, store it in the secret store, and revoke it afterward.

A long-term Bedrock API key supports bearer-token authentication and can remain valid throughout the 72-hour run without renewal. A dedicated IAM user with narrowly scoped permissions, encrypted storage, and revocation after the evaluation meets the stated security requirements.

DOP-C02

A company's security monitoring system detects that an IAM user's access key was published in a public repository. AWS CloudTrail records show that an attacker used the key to call sts:GetSessionToken successfully. The user has not assumed any IAM roles. A DevOps engineer deactivates all of the user's access keys and records a containment timestamp after those actions complete.

The engineer must now deny requests made with the user's temporary credentials issued before the containment timestamp. The IAM user must be retained, and temporary credentials issued after the timestamp through an approved recovery process must remain usable with the user's existing permissions.

Which additional action meets these requirements?

Show answerCORRECT ANSWER

Attach an inline policy to the IAM user that explicitly denies all actions on all resources when aws:TokenIssueTime is earlier than the containment timestamp.

Temporary credential requests include aws:TokenIssueTime. An explicit deny using DateLessThan against the containment timestamp blocks the user's previously issued temporary credentials, including those obtained through GetSessionToken. Later sessions do not match the condition and remain subject to the user's existing permissions. Deactivating the access keys separately blocks requests signed directly with those keys.

DEA-C01

An AWS Glue job retrieves a database password from AWS Secrets Manager in us-east-1. The secret is encrypted with a dedicated customer managed AWS KMS key. The job's execution role already allows secretsmanager:GetSecretValue on the secret, but it has no KMS permissions. The key policy enables the account to delegate permissions through IAM policies.

Security requires the role to decrypt the password through Secrets Manager, but prohibits the role from calling KMS directly to decrypt ciphertext, even if the role obtains the ciphertext and its encryption context. No other policies grant or restrict the role's KMS access.

Which permission should a data engineer add to the execution role to meet these requirements?

Show answerCORRECT ANSWER

Allow kms:Decrypt on the key ARN with a StringEquals condition requiring kms:ViaService to equal secretsmanager.us-east-1.amazonaws.com.

Secrets Manager calls KMS on behalf of the execution role when retrieving the encrypted secret. The kms:ViaService condition permits this service-mediated decryption while excluding direct KMS calls from the role.

Practice Credential Management in a full exam

Questions on Credential Management in each practice exam, with every answer option and an explanation for each:

Start a free practice exam

What is Credential Management

3 min read

Credential management governs how authentication material is issued, protected, used, replaced, and retired. In cloud environments, this includes credentials for engineers, administrators, applications, automation pipelines, and emergency access. Authentication establishes who is making a request; authorization determines what that identity may do. A securely stored credential can still be dangerous if its identity has excessive permissions, so credential management must work alongside least-privilege access policies.

Different credentials serve different purposes. An AWS IAM user can have a console password and, separately, access keys for programmatic requests. An access key consists of an access key ID and a secret access key. Temporary AWS credentials also include a session token and expire automatically. Changing a console password does not rotate access keys, and deleting an access key does not remove console access. Engineers must inventory and manage each authentication method independently.

Prefer temporary credentials over long-lived access keys wherever possible. A workload running on an AWS compute service can obtain role credentials through its supported credential mechanism rather than embedding keys in code. Developers can use federated sign-in and role sessions instead of maintaining individual long-lived keys. AWS SDK credential providers can retrieve and refresh role credentials automatically. Temporary credentials reduce persistent exposure, but they remain sensitive and can be abused until they expire or access is otherwise restricted.

When an application needs a database password or external API token, store it in a dedicated secret store rather than source code, container images, shared documents, or logs. For example, a workload can use its IAM role to retrieve a database password from AWS Secrets Manager. AWS Systems Manager Parameter Store also supports encrypted SecureString parameters. Secrets Manager provides managed rotation capabilities for supported integrations and customizable rotation workflows; Parameter Store does not provide equivalent built-in secret rotation. Encryption at rest alone is insufficient: restrict retrieval permissions, protect decryption access, and prevent applications from exposing retrieved values.

Password policies establish requirements such as minimum length, complexity, and reuse restrictions. AWS IAM account password policies apply to IAM user console passwords, not root user passwords, access keys, or passwords managed by an external identity provider. Configure workforce password controls in the system that actually authenticates those users. Use unique passwords and multifactor authentication rather than relying on complexity alone. Enabling MFA for console sign-in does not automatically require MFA for every request signed with an IAM user's long-lived access keys.

Root user protection deserves separate controls because root has highly privileged account capabilities. Do not use root for everyday administration, and do not create root access keys. Protect the root password with a strong, unique value, enable MFA, and secure the associated email account and recovery channels. Restrict access to root credentials, monitor root activity, and reserve root sign-in for tasks that require it. Where AWS Organizations centralized root access management is configured, eligible member accounts can have long-term root credentials removed; the management account still requires its own root protection.

Rotation and revocation must be operationally safe. When replacing a long-lived key, update its consumers, verify that they use the replacement, and then deactivate and delete the old key. For a leaked credential, revoke or disable it promptly and investigate its use; removing it from a repository does not invalidate it. Secret retrieval adds cost, latency, and availability dependencies, while caching reduces requests but can delay adoption of rotated values. Design refresh behavior, monitoring, and incident procedures around these trade-offs.

Learn Credential Management in lessons

6 short lessons on Credential Management in 5 certification courses.

Open the lessons