Oracle 1Z0-997-26 Exam Questions & Answers (November 2026)
Oracle Cloud Infrastructure Architect Professional exam questions with 14 free practice questions and explained answers. Updated November 21, 2026. · See all Oracle exams
Answer 14 scenario questions on this page and see your score as you go. No sign-up needed.
Exam Details
- Exam Code
- 1Z0-997-26
- Vendor
- Oracle
- Questions
- 50
- Duration (minutes)
- 90
- Passing Score
- 68%
- Question Types
- Single Choice
- Official Page
- View Exam Page
Architect High Availability and Disaster Recovery Solutions
15% of the examThis topic covers designing OCI workloads to remain available during component or regional failures. Candidates must choose suitable high-availability patterns and configure Full Stack Disaster Recovery to coordinate recovery across regions.
A customer-facing application runs in an OCI region with three availability domains. Its compute instances currently reside in one availability domain and serve requests through an OCI Load Balancer. The business requires the application to continue serving traffic if that entire availability domain becomes unavailable. The team wants the load balancer to route around failed instances without requiring an operator to launch replacements.
Which design best meets the availability requirement?
Show Answer
Correct Answer
A
Option A is correct because distributing the instance pool across availability domains lets the regional load balancer direct traffic to instances outside a failed domain. Option B is not correct because fault domains protect against failures within an availability domain, not its loss. Option C does not meet the regional recovery requirement. Option D cannot place one instance across multiple availability domains.
A service runs in one OCI region and has a prepared recovery environment in a second region. Its application compute instances and required storage are supported for Full Stack Disaster Recovery, and the recovery environment must be brought online in a coordinated operation. The operations team needs a planned switchover for maintenance and a separate recovery path for an unexpected regional outage. It wants to manage execution through OCI rather than maintain custom orchestration scripts.
Which design best supports the required regional disaster recovery operations?
Show Answer
Correct Answer
A
Option A is correct because paired primary and standby DR protection groups let Full Stack Disaster Recovery coordinate supported stack resources through DR plans for both planned and unplanned recovery. Option B lacks the peer protection group required for the cross-region relationship. Option C provides an alert and a compute action, not coordinated stack recovery. Option D depends on custom scripts rather than FSDR plans.
Architect Cloud-Native Solutions
20% of the examThis topic covers the design and deployment of cloud-native applications on OCI using containers, microservices, and serverless services. Candidates must choose suitable OCI services and use DevOps and Infrastructure as Code practices to build, deploy, and operate these applications.
A retail team runs a containerized ordering service on an existing OKE cluster and stores its images in OCIR. The team wants every approved code change to build and test a new image, publish that image to the registry, and roll out a Kubernetes manifest to the existing cluster. They want OCI DevOps to manage the release flow, with no operator running deployment commands.
Which deployment design best meets the team’s requirements?
Show Answer
Correct Answer
A
Option A is correct because the build pipeline can publish the image to OCIR and pass the Kubernetes manifest to an OKE deployment stage for release to the existing cluster. Option B creates unnecessary clusters and uses Terraform for application rollout. Option C uses Functions rather than OKE. Option D leaves the required deployment step manual.
A finance department has a small internal reporting service packaged as one Linux container. It must remain available at a private endpoint during business hours, but its demand is modest and stable. The service does not need Kubernetes APIs, coordinated multi-container scheduling, or cluster-level scaling. The operations team also does not want to patch or administer virtual machine hosts.
Which OCI compute service is the best fit for this workload?
Show Answer
Correct Answer
B
Option B is correct because the service needs to run as a container but does not need Kubernetes orchestration or host management. Option A adds a cluster that the team does not require. Option C is not the right fit for a continuously running container service. Option D makes the team responsible for managing the host and container runtime.
A development group uses Terraform to define an Oracle Functions application and function, with the function image stored in OCIR. The group must deploy the same design to development and production, changing the image URI between releases. The platform team requires each change to have a reviewable plan and wants Terraform state managed by an OCI service rather than kept on individual workstations.
Which approach best provides repeatable deployments and managed Terraform state?
Show Answer
Correct Answer
A
Option A is correct because Resource Manager runs Terraform for the stack and manages its state, while the image URI can be supplied as a stack input. Option B leaves state management to a shared file and workstation process. Option C treats Terraform as a Kubernetes manifest. Option D provides no repeatable, state-managed deployment.
Architect Security Solutions
25% of the examThis topic covers layered OCI security controls for administrative access, network traffic, and public applications. Candidates must select services and policies that enforce access requirements while preserving the organization’s existing network controls.
A financial services team operates Linux compute instances in private subnets. The instances have no public IP addresses, and policy prohibits assigning them one. Administrators connect from a fixed corporate CIDR, need temporary SSH access for maintenance, and must not share a permanent jump-host account. The team also wants a control that limits how long an access session remains available.
Which design best meets the access requirements while keeping administrative access controlled?
Show Answer
Correct Answer
A
Option A is correct because OCI Bastion provides time-limited managed SSH sessions to private compute targets and can restrict session creation to the approved administrator CIDR, so no target needs a public IP. Option B is incorrect because it exposes SSH on public addresses. Option C is incorrect because Network Firewall does not broker short-lived administrator sessions. Option D is incorrect because WAF protects HTTP applications, not SSH access.
A company hosts a public ordering application behind an OCI Load Balancer. The application team has observed repeated HTTP requests that appear to target common web vulnerabilities, along with bursts of requests to a small set of endpoints. The team wants centrally managed protection before requests reach the application and does not want to change application code or rely only on IP-based network rules.
Which configuration best addresses the application-layer risks and request volume requirement?
Show Answer
Correct Answer
C
Option C is correct because an OCI WAF policy can be associated with the public load balancer and apply HTTP protection rules and request-rate controls before traffic reaches the application. Option A is incorrect because Network Firewall does not provide WAF managed HTTP exploit rules. Option B is incorrect because NSGs filter network flows, not HTTP request signatures. Option D is incorrect because certificates provide TLS identity, not request filtering.
An enterprise runs application and payment services in private VCN subnets. Several teams share those subnets, and IP ranges may change during network maintenance. Security policy requires the application workload to reach a specific payment service, while other workloads must not gain access simply because they share a subnet or address range. Existing routing and network security controls must remain in place.
Which approach best supports the requested workload-based communication controls?
Show Answer
Correct Answer
B
Option B is correct because ZPR uses resource security attributes and policies to express which labeled workloads may communicate, while routing and existing network controls still provide reachability. Option A is incorrect because CIDR rules grant access by address and do not express workload identity. Option C is incorrect because WAF handles web requests, not east-west service access. Option D is incorrect because subnet placement alone does not establish an attribute-based communication policy.
Architecting, Implementing, and Operating Databases in OCI
20% of the examThis topic covers database design and operation with OCI Base Database Service and Autonomous AI Database, including capacity, protection, and data placement choices. Candidates must select database architectures and service features that meet workload, recovery, and regional access requirements.
A retail company runs its order application on Autonomous AI Database. Most days, demand is steady, but short promotions create sharp workload peaks. The team wants the database to add compute capacity during those peaks and reduce it afterward. It does not want to pay for peak capacity throughout the month, and the application can use the existing database service without changes.
Which change best meets the application team’s capacity requirement?
Show Answer
Correct Answer
A
Option A is correct because the baseline covers ordinary demand while compute auto scaling can add capacity during the stated bursts without holding peak OCPUs all day. Option B is not correct because permanent peak provisioning wastes capacity during normal periods. Option C is not correct because regional read replicas do not address burst capacity for this database. Option D is not correct because storage auto scaling adds storage, not compute.
A finance team operates an Oracle Database on OCI Base Database Service. Its audit policy requires automated backups and the ability to recover the database to a selected point within a defined retention window. The database team wants managed protection rather than scheduled scripts maintained on the database hosts. A standby database is already used for a separate availability requirement, but the team needs a distinct backup and recovery design.
Which backup design best meets the database team’s requirements?
Show Answer
Correct Answer
A
Option A is correct because the protection policy provides managed database backups and a defined recovery window for the Base Database Service database. Option B is not correct because boot volume backups do not replace database backup and recovery management. Option C is not correct because a standby is not a backup retention solution. Option D is not correct because occasional exports cannot provide the required point-in-time recovery.
An enterprise application serves customers in Europe and the Americas from regional application stacks. Most transactions affect records for one customer, and cross-region transactions are uncommon. The company wants one logically managed database service, with customer data kept together and routine requests directed to the region that owns those records. It cannot accept independently writable regional copies that may diverge.
Which database architecture best supports the application’s locality and data ownership requirements?
Show Answer
Correct Answer
A
Option A is correct because the shard key can group customer records and support routing to the shard that owns the data, reducing the distance for regional transactions. Option B is not correct because a single central database adds network distance for remote users. Option C is not correct because independent writable copies create conflicting data ownership. Option D is not correct because OCI Cache is not the authoritative database for durable transactions.
Implementing Observability Solutions
20% of the examThis topic covers monitoring OCI resources, collecting logs, and routing observability data for investigation. Candidates must be able to configure metric alarms, select suitable log sources, and use OCI services to discover and analyze resources.
A service team runs 60 compute instances in one compartment. CPU utilization varies widely because batch workers and API servers share the fleet, and the team has missed incidents when one instance was saturated while the fleet average stayed low. The on-call rotation already uses an OCI Notifications topic and requires an actionable alert for each affected instance.
Which alarm design best meets the team's requirement?
Show Answer
Correct Answer
A
Option A is correct because grouping the CPU metric by resourceId keeps each instance as a separate metric stream, so one busy instance is not hidden by a fleet average, and the alarm can publish to the Notifications topic. Option B still combines instances. Option C uses Logging rather than metric alarm evaluation. Option D provides a chart but sends no alarm notification.
A support team operates a compute-based application and investigates intermittent request failures. The application writes useful error details to a local file, but only OCI service logs are currently enabled and searchable. The team needs to retain those service logs, collect the application file centrally, and run time-bounded investigations across both sources without moving log data to a separate tool.
Which change best provides the requested log collection and investigation workflow?
Show Answer
Correct Answer
A
Option A is correct because a custom log and Logging agent configuration collect the application file, while the existing service logs remain available for search. Option B is not correct because Audit records OCI API activity, not local file contents. Option C is not correct because service logs do not collect arbitrary compute files. Option D confuses log records with metric data.
A monitoring team is onboarding a new application and must verify that every compute instance tagged for the application is included in its monitoring plan. The instances are spread across several compartments, and some have not yet produced metrics or logs because they are newly created. The team needs a resource inventory that does not depend on observed telemetry.
Which approach best identifies the instances that need to be added to the monitoring inventory?
Show Answer
Correct Answer
A
Option A is correct because OCI Search can discover matching compute resources across the selected compartments using resource attributes and tags, including instances that have not yet emitted metrics or logs. Option B omits instances without metric data. Option C only finds log producers. Option D routes log data but does not inventory compute resources.
Continue Your 1Z0-997-26 Preparation
You have seen 14 questions from a full set of 100. Prepare with the 1Z0-997-26 dumps PDF and the Web-Based Practice Test.
- Compatible with All Devices
- Printable Format
- No Download Limits
Practice Test
- Check your timing and weak areas
- No Practice Limits
- Regularly Updated
- 100% Money-Back Guarantee
- 90 Days Free Updates
- Instant Access
Frequently Asked Questions
Which part of 1Z0-997-26 tends to demand the most careful architectural judgment?
Security accounts for 25% of the exam and asks you to think across administrative access, network traffic, and public applications. The challenge is choosing layered controls that satisfy access requirements while preserving an organization’s existing network controls. Memorizing service names is not enough if you cannot connect them to the stated constraints. Practise by outlining the required access, traffic path, and policy controls for a sample OCI design before selecting services.
How does the 90-minute limit affect the way I should approach the questions?
The exam has 50 questions in 90 minutes, so you have limited time to reason through scenarios spanning five substantial topic areas. Cloud-native architecture, security, databases, observability, and recovery can each involve several service choices. Avoid spending too long debating one design detail. Practise making a defensible choice from the requirements, then move on and return to difficult items if time allows.
Why can studying OCI services individually leave gaps in preparation?
The exam tests architecture and operation decisions, not just definitions. A question about a cloud-native workload may involve containers, serverless services, DevOps, or Infrastructure as Code. A recovery design may require matching availability patterns with Full Stack Disaster Recovery across regions. Build study notes around requirements and trade-offs, then practise explaining why a service choice fits the workload rather than learning each topic in isolation.
What can I use when I need to review but cannot work in an OCI lab?
Some preparation time will be spent away from a console, and that can make it difficult to keep architecture decisions fresh. The Exam4Success PDF guide is useful for reviewing topic explanations away from a lab, particularly for comparing recovery, security, and database design choices. Use it to identify areas that need hands-on follow-up, then return to an OCI environment when you can to practise configuring or tracing those designs.
What practical capabilities does this professional architect exam assess?
It assesses your ability to design and operate OCI solutions across availability, cloud-native applications, security, databases, and observability. You should be able to select high-availability patterns, plan regional recovery, choose database features for workload and access needs, and configure monitoring and logs for investigation. Treat the exam as a check of architectural decision-making across connected services. A useful preparation goal is to explain a complete design and its operational needs.
Will preparing for this exam help me think beyond application deployment?
The topic coverage extends from deployment into the controls and operations needed to run OCI workloads. Alongside containers, microservices, serverless services, DevOps, and Infrastructure as Code, the exam covers layered security, database protection, metric alarms, and log routing. That breadth is useful for candidates who want to connect application architecture with ongoing operations. As you study, map each workload design to its access controls, recovery approach, and observability plan.
What should I study first if I am setting up a plan for 1Z0-997-26?
Start with security, which represents 25% of the exam, then give substantial time to cloud-native solutions, databases, and observability, each at 20%. High availability and disaster recovery represent 15%. These weights can help you allocate review time, but they should not replace coverage of any topic. Begin by assessing your confidence in each area, then spend the earliest sessions on the weakest topics with significant exam weight.
How can I tell if I am ready to handle the exam’s pace and topic range?
Use a timed Exam4Success Practice Test to check whether you can sustain your decision-making across varied OCI topics within a fixed session. Review missed and uncertain answers by topic, then distinguish knowledge gaps from rushed reading or difficulty interpreting requirements. The test can help identify weak areas, but it is not a guarantee of an exam result. Set a specific follow-up review for the two topics that need the most attention.
When should I focus my database preparation on service selection rather than recovery?
Prepare both, because the database topic covers OCI Base Database Service and Autonomous AI Database as well as capacity, protection, data placement, workload needs, recovery, and regional access. Begin by listing what a workload requires, then compare how the relevant service and architecture choices address those needs. After that, review protection and regional access as part of the same design. This makes it easier to apply database knowledge to scenario-based decisions.
Exam4Success is an independent exam preparation provider and is not affiliated with or endorsed by Oracle.