# Patch Management

Patch management is the controlled process of identifying, testing, deploying, and verifying software updates to reduce vulnerabilities and maintain reliable systems.

Cloud concept. Updated 2026-10-29. Page: https://blowstack.com/tools/cloud-concepts-hub/patch-management

> Patch management keeps operating systems, applications, and dependencies updated through repeatable policies and deployment workflows. In cloud environments, responsibility depends on the service model: providers patch infrastructure they operate, while customers patch the software layers they control. Effective patching combines automation with risk-based prioritization, availability safeguards, and continuous validation.

Practice exams covering Patch Management: [CLF-C02 AWS Certified Cloud Practitioner](https://blowstack.com/practice-exams/aws-certified-cloud-practitioner), [DOP-C02 AWS Certified DevOps Engineer - Professional](https://blowstack.com/practice-exams/aws-certified-devops-engineer-professional), [SCS-C03 AWS Certified Security - Specialty](https://blowstack.com/practice-exams/aws-certified-security-specialty), [SAP-C03 AWS Certified Solutions Architect - Professional](https://blowstack.com/practice-exams/aws-certified-solutions-architect-professional).

## 6 key aspects of Patch Management

Effective patch management combines clear ownership, risk-based policy, automated deployment, availability controls, and continuous verification.

1. **Layer-specific responsibility**: Assign patch ownership according to who controls each software layer. Customers patch EC2 guest operating systems and installed software, while AWS patches the underlying infrastructure it operates.
2. **Risk-based patch policy**: Define approved updates, remediation deadlines, and exception procedures. Prioritize actively exploited vulnerabilities and exposed critical workloads rather than relying only on severity ratings.
3. **Automated, repeatable deployment**: Use inventory, resource targeting, schedules, and controlled execution to deploy updates consistently. Monitor agent health, permissions, and repository access so failed or missing assessments do not create false confidence.
4. **Availability and recovery safeguards**: Test updates before broad deployment and use staged rollouts with concurrency limits. Account for restarts, spare capacity, backups, and recovery options before changing production systems.
5. **In-place and immutable remediation**: Choose between updating running resources and replacing them with newly built, patched images. Keep deployment templates and image references current to prevent vulnerable versions from returning during scaling or recovery.
6. **Continuous verification and evidence**: Combine patch compliance checks, vulnerability reassessment, and workload health tests. Retain results and time-limited exceptions because successful installation does not by itself establish security or application correctness.

## Cloud services for Patch Management

Systems Manager provides patch assessment and installation for managed nodes, while Amazon Inspector supplies vulnerability findings that help prioritize work and validate remediation within supported coverage. EC2 Image Builder supports an alternative workflow: build and test updated images, then replace workloads through deployment automation. AWS Config can evaluate patch compliance using Systems Manager data and flag resources requiring attention. Together, these services support a feedback loop from discovery through remediation to evidence, but engineers must still define ownership, approval policies, rollout safeguards, application tests, and exception handling.

- **[AWS Systems Manager](https://blowstack.com/tools/cloud-services-hub/aws-systems-manager)**: Patch Manager, a capability of AWS Systems Manager, scans managed nodes for patch compliance and installs approved patches using patch baselines. Patch policies and Maintenance Windows support scheduled execution, while managed-node connectivity, permissions, and package-repository access remain operational prerequisites.
- **[Amazon Inspector](https://blowstack.com/tools/cloud-services-hub/amazon-inspector)**: Amazon Inspector identifies software vulnerabilities in supported EC2 instances, container images in Amazon ECR, and Lambda functions within its enabled scanning coverage. Use findings to prioritize remediation and reassess exposure afterward; Inspector does not deploy patches.
- **[EC2 Image Builder](https://blowstack.com/tools/cloud-services-hub/ec2-image-builder)**: EC2 Image Builder automates the creation, testing, and distribution of machine images and container images, including update steps configured in build workflows. Deploying those images through a separate rollout process supports immutable patching; existing workloads remain unchanged until replaced.
- **[AWS Config](https://blowstack.com/tools/cloud-services-hub/aws-config)**: AWS Config can evaluate managed-instance patch compliance using a managed rule that relies on Systems Manager compliance information. This supports compliance tracking and remediation workflows but does not independently scan for vulnerabilities or install updates.

## Patch Management exam questions

Real practice exam questions on Patch Management, with the correct answer and its explanation.

### Question 1 (CLF-C02)

**Question:**

A company runs applications on Amazon EC2 instances. Its security policy requires approved guest operating system patches to be installed during a weekly maintenance window. The company wants to automate this work while following the AWS shared responsibility model. Which approach should it take?

**Correct answer:** Use AWS Systems Manager Patch Manager with approved patch baselines and a weekly maintenance window, while retaining responsibility for verifying patch compliance.

**Explanation:** Patch Manager can automate installation of approved guest operating system patches on managed EC2 instances. The customer remains responsible for its patch policy and verifying results, while AWS patches the underlying infrastructure it operates.

### Question 2 (SAP-C03)

**Question:**

A company runs a production application on 60 Amazon EC2 instances behind an Application Load Balancer in one AWS Region. All instances are Systems Manager managed nodes. The application can meet its availability requirements with 59 instances, but not with 58. The fleet remains unchanged during maintenance.

The operations team must automate installation of approved operating system patches. For each instance, the process must deregister the instance from the target group, wait for connection draining to finish, install patches and perform any required reboot, wait for Systems Manager connectivity, validate the application locally, and then register the instance again. If patching or validation fails, the instance must remain deregistered, and processing must stop before another instance is deregistered.

Which solution meets these requirements with the LEAST custom orchestration code?

**Correct answer:** Create a Systems Manager Automation runbook that performs the complete procedure for one instance, including explicit draining and readiness checks. Use rate-controlled Automation targeting for the fleet with maximum concurrency set to 1 and maximum errors set to 0. Configure patching and validation failures to fail that instance's execution without running the registration step.

**Explanation:** Rate-controlled Automation can apply the entire runbook to one target at a time. A concurrency of 1 prevents another instance from entering maintenance while the current instance's procedure is running. An error threshold of 0 stops subsequent target executions after the first failure. Explicit waits handle draining and reboot recovery, and failure paths that bypass registration keep an unsuccessfully maintained instance out of service.

### Question 3 (SCS-C03)

**Question:**

A company uses AWS Systems Manager Patch Manager to update Red Hat Enterprise Linux EC2 instances. A custom patch baseline approves security updates and explicitly rejects a package version that causes application failures. During a pilot deployment, an approved security update installs the rejected package version as a dependency. Amazon Inspector subsequently confirms that the vulnerability addressed by the security update is resolved, but application tests fail.

The remaining instances have not received either update. Company policy prohibits installing the rejected package version, including as a dependency. If an approved security update cannot be installed without that version, the update must remain unapplied and be escalated for an alternative remediation. The company also requires continued reporting of unresolved vulnerabilities.

Which change should a security engineer make before continuing the automated rollout?

**Correct answer:** Keep the package version in the baseline's rejected patches list and set the rejected-patch action to BLOCK. Continue scheduled installation and compliance scans, retain Inspector assessment, and escalate updates blocked by the rejected dependency.

**Explanation:** BLOCK prevents explicitly rejected patches from being installed, including as dependencies, and prevents installation of patches that require them. Scheduled patch compliance scans and Inspector assessments provide visibility into unapplied updates and unresolved vulnerabilities. The blocked security update must follow the company's alternative-remediation process.

### Question 4 (DOP-C02)

**Question:**

A company manages 2,500 Windows Server EC2 instances with AWS Systems Manager. All instances have the tag Patch Group=Production, which is registered with a custom patch baseline. The baseline automatically approves Critical and Important security updates 14 days after release. A weekly maintenance window targets instances by their Environment=Production tag.

A security team identifies an actively exploited vulnerability on 200 instances. A specific Windows KB update released 2 days ago fixes the vulnerability and has passed application testing. The affected instances also have the tag EmergencyPatch=Required.

A DevOps engineer must immediately install the KB update and complete any required reboots on the affected instances. No other update released within the last 14 days may become approved. Unaffected instances must retain their existing patch approval policy. Future weekly patching must continue to include all production instances.

Which solution meets these requirements?

**Correct answer:** Copy the production baseline into an emergency baseline, retain its existing approval rules, and add the KB update to its explicitly approved patches. Register the emergency baseline with a new patch group and change the affected instances' Patch Group tag to that group. Run AWS-RunPatchBaseline with Operation=Install and RebootOption=RebootIfNeeded, targeting EmergencyPatch=Required.

**Explanation:** The separate baseline explicitly approves only the emergency KB in addition to the existing approval policy. Assigning the affected instances to its registered patch group makes Patch Manager select that baseline. The Install operation applies approved missing updates and completes required reboots. The original baseline remains unchanged for unaffected instances, and the Environment-based maintenance window continues to include both patch groups.

## Practice Patch Management in a full exam

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

- [SCS-C03](https://blowstack.com/practice-exams/aws-certified-security-specialty): 6 questions
- [DOP-C02](https://blowstack.com/practice-exams/aws-certified-devops-engineer-professional): 3 questions
- [CLF-C02](https://blowstack.com/practice-exams/aws-certified-cloud-practitioner): 1 question
- [SAP-C03](https://blowstack.com/practice-exams/aws-certified-solutions-architect-professional): 1 question

## What is Patch Management

Patch management is a lifecycle, not simply the act of installing updates. Engineers inventory software, identify applicable fixes, assess urgency, test compatibility, deploy approved updates, and verify the resulting state. Security patches address known vulnerabilities, while other patches can resolve reliability defects. A practical policy defines remediation deadlines, maintenance schedules, approval rules, exception handling, and the evidence required to demonstrate completion.

In the cloud shared responsibility model, patching is a shared control applied to different layers, rather than a task that both parties perform on the same components. AWS patches the physical infrastructure and virtualization layers that it operates. For an Amazon EC2 instance, the customer is responsible for patching the guest operating system and customer-installed applications. With managed services, AWS operates and patches more of the stack, but customers still manage applicable maintenance settings, supported versions, and their own code and dependencies. Using a managed service does not eliminate every customer update obligation.

Start with an accurate inventory and ownership model. Classify workloads by operating system, environment, business criticality, exposure, and maintenance requirements. Prioritize vulnerabilities using exploitability, active exploitation, asset importance, and network reachability rather than severity scores alone. An internet-facing server with an actively exploited vulnerability may require an emergency rollout, while a lower-risk update can follow a normal testing cycle. Exceptions should have an owner, an expiration date, and compensating controls.

Automation translates policy into repeatable execution. For example, an operations team can use AWS Systems Manager Patch Manager to scan managed EC2 instances against patch baselines and install approved patches on a schedule. Baselines determine which updates are approved; they are not substitutes for application compatibility testing. Successful execution also depends on supported operating systems, healthy management agents, suitable permissions, and connectivity to management endpoints and package repositories. An unreachable or unmanaged instance must not be mistaken for a compliant instance.

Deploy updates progressively through development, staging, and production. Limit concurrency and stop or investigate when failures exceed defined thresholds. For a load-balanced application, drain and patch a subset of instances while preserving enough healthy capacity. Some updates require a reboot or service restart before the fix becomes effective, so a successful installation command alone does not prove remediation. Prepare backups or replacement capacity and a recovery plan; uninstalling a patch is not always supported or safe.

Immutable deployment is an alternative to modifying running servers. Build and test a patched machine image or container image, deploy replacement workloads, and retire the old versions. Updating an image does not patch instances or containers already running from it. Container users remain responsible for updating image contents and application dependencies, even when the provider manages the underlying compute hosts. Ensure scaling configurations reference the approved image so newly launched capacity does not reintroduce vulnerable software.

Continuous validation closes the loop. Patch compliance reports show whether resources satisfy an approval baseline, while vulnerability scanning identifies known weaknesses within its supported coverage. Amazon Inspector can identify vulnerabilities on supported workloads, but it does not install their patches. After remediation, repeat scans, confirm relevant restarts, run application health checks, and investigate remaining findings. A resource can satisfy its patch baseline yet still contain a vulnerability outside that baseline or scanning scope. Preserve deployment results, exceptions, and validation evidence to support audits and detect drift.

## Learn Patch Management in lessons

5 short lessons on Patch Management in 4 certification courses.

- CLF-C02: Patch Management: responsibility across cloud service layers (Premium), https://app.blowstack.com/lessons/patch-management-responsibility-across-cloud-service-layers-clf-c02
- SAP-C03: Patch Management: automated patch rollouts (Premium), https://app.blowstack.com/lessons/patch-management-automated-patch-rollouts-sap-c03
- SCS-C03: Patch Management: controlled automated rollouts (Premium), https://app.blowstack.com/lessons/patch-management-controlled-automated-rollouts-scs-c03
- DOP-C02: Patch Management: automated fleet patching (Premium), https://app.blowstack.com/lessons/patch-management-automated-fleet-patching-dop-c02
- SCS-C03: Patch Management: verifying effective remediation (Premium), https://app.blowstack.com/lessons/patch-management-verifying-effective-remediation-scs-c03
