Select Two (Multi-Select) Questions

Grouping and filtering costs using OCI Defined Tags

Cost Analysis supports grouping and filtering spend by OCI defined tags, letting administrators break costs down by the tag keys applied to resources and compartments, which satisfies the reporting and analysis requirement in the stem.

Visualizing cumulative cost trends over customizable time ranges (e.g., daily, monthly)

Cost Analysis provides cumulative cost trend charts, letting administrators aggregate spend across customisable daily or monthly ranges. This satisfies the requirement to visualise spending patterns over time, supporting trend analysis and reporting without exporting data to external tools.

Exporting cost charts and tabular data to CSV format

Exporting cost charts and tabular data to CSV satisfies the reporting requirement directly: Cost Analysis generates downloadable CSV files of both visualisations and underlying tabular figures, letting administrators circulate spend data to stakeholders outside OCI. This native export capability is one of the three supported features the stem requests.

Consolidate multiple non-CDB databases into a single Multitenant Container Database (CDB) with multiple PDBs.

Consolidating non-CDBs into a single Multitenant CDB with multiple PDBs directly addresses the resource-utilisation constraint: PDBs share one instance's SGA, background processes and CPU, eliminating per-database overhead. This raises consolidation density on the same Exadata infrastructure, so fewer dedicated resources are provisioned, lowering cost while maintaining performance isolation.

Configure Exadata Smart Flash Cache policies appropriately for OLTP versus DSS workloads.

Smart Flash Cache policies let you prioritise columnar flash for DSS scans or keep OLTP hot blocks cached, matching flash behaviour to each workload's access pattern. Correct tuning reduces physical I/O and latency, improving performance without adding infrastructure cost.

Dynamically scale OCPU counts on the Exadata infrastructure based on peak and off-peak workload schedules.

Elastic OCPU scaling on Exadata infrastructure lets you reduce allocated compute during off-peak windows and restore it for peak demand, so you pay only for capacity actually needed. This directly satisfies the cost-efficiency and resource-utilisation constraint by matching provisioned OCPUs to the fluctuating workload schedules of multiple large databases.

Deploying Windows Server instances on OCI Compute attached to high-performance OCI Block Volumes to act as native SMB file servers.

Windows Server on OCI Compute with attached Block Volumes provides native SMB shares, preserving NTFS permissions and Active Directory integration. This satisfies the SMB protocol constraint without re-architecting applications, unlike NFS-based File Storage or Object Storage.

Evaluating OCI Marketplace partner solutions that provide enterprise NAS appliances with native SMB protocol support on OCI.

OCI File Storage Service natively speaks NFSv3 only, so Windows SMB shares need another route. Marketplace partner NAS appliances provide native SMB protocol support on OCI, preserving Windows ACLs and SMB semantics without gateways or re-architecture.

Standard OCI File Storage Service (FSS) natively speaks NFSv3 and does not support SMB/CIFS natively without third-party gateways or workarounds.

This is the constraint driving the whole decision: OCI File Storage Service natively speaks NFSv3 and offers no native SMB/CIFS support. Windows file shares therefore require third-party gateways, workarounds, or an alternative storage architecture that does provide SMB.

Deploying application tiers across two OCI regions with DNS Traffic Management health checks for automated failover

Multi-region deployment with DNS Traffic Management health checks provides automated failover: when the primary region's endpoints fail health probes, traffic is steered to the secondary region. This directly mitigates a full regional outage, which single-region fault domain or availability domain designs cannot survive.

Replicating critical block volumes and object storage data to the secondary region

Cross-region replication of block volumes and object storage ensures data survives a regional outage and can be restored or attached in the secondary region. Without replicated data, compute failover alone leaves the application without its state, so this underpins the recovery strategy.

Configuring Oracle Data Guard between the primary database and a standby database in a secondary region

Oracle Data Guard replicates redo from the primary database to a standby in a secondary region, enabling failover with minimal data loss when the primary region becomes unavailable. This addresses the database tier specifically, complementing compute and storage resilience elsewhere.

Automated and policy-driven backups to OCI Object Storage with cross-region replication

Policy-driven backups replicated cross-region to OCI Object Storage give an independent recovery copy that survives regional loss, underpinning both uptime and disaster recovery. It satisfies the recovery objective by enabling restore when primary database copies are unavailable.

Oracle Data Guard standby database in a remote region for disaster recovery

A Data Guard standby in a remote region maintains a synchronised replica that can be activated after a regional outage, providing the disaster recovery capability the architecture demands. It directly satisfies the requirement for optimal recovery and uptime across regions.

Oracle Real Application Clusters (RAC) across fault domains or availability domains for local HA

RAC instances span fault domains or availability domains, so a single node or AD failure does not stop database service. This satisfies the local high-availability requirement by providing instance-level redundancy and automatic failover within the region, complementing Data Guard for disaster recovery.

Implementing Conflict Detection and Resolution (CDR) rules for active-active multi-master topologies

Active-active multi-master topologies let the same row be updated at multiple nodes simultaneously, so CDR rules detect and resolve those conflicting changes. Without them, replication applies divergent updates and data diverges, breaking the load-distribution design.

Configuring unique primary key generation strategies (e.g., sequence offsets) to prevent key collisions

Unique sequence offsets per database prevent duplicate primary keys when GoldenGate replicates inserts across sites, satisfying the load-distribution constraint where multiple databases accept writes concurrently. Without distinct ranges, bidirectional replication triggers key collisions that halt apply processes, so partitioning the key space is essential for conflict-free convergence.

Managing DDL replication synchronization to ensure schema modifications match across nodes

GoldenGate replicates DML by default, so DDL executed on one node is not automatically applied elsewhere. Coordinating DDL replication keeps schemas identical across nodes, preventing replication errors when table structures diverge in the multi-database topology.

Distributed Tracing across microservices

Distributed tracing follows a request across microservice boundaries, capturing span timing and causal relationships to pinpoint latency bottlenecks. This satisfies the stem's requirement for an APM capability, since OCI APM natively ingests trace data from instrumented services, unlike infrastructure monitoring tools that report only host-level metrics.

Real User Monitoring (RUM) to track end-user browser performance

Real User Monitoring captures browser-side telemetry—page load timings, JavaScript errors and AJAX calls—via an injected agent, satisfying the stem's requirement for an APM feature. OCI APM includes RUM natively, alongside distributed tracing and synthetic monitoring, so end-user experience data flows into the same service without third-party tooling.

Synthetic Monitoring using browser and script-based tests

Synthetic Monitoring runs scheduled browser and script-based tests from external vantage points, simulating user journeys to detect availability and performance regressions proactively. This is a built-in APM feature, distinct from RUM, which measures real traffic rather than scripted probes.

Distribute OKE worker node pools across multiple Fault Domains or Availability Domains.

Spreading worker node pools across multiple Fault Domains or Availability Domains ensures that a single domain failure cannot take down all nodes hosting the microservices, satisfying the high-availability and resilience requirement by removing a shared failure domain.

Configure Kubernetes Pod Disruption Budgets (PDBs) to control minimum available pods during maintenance or scaling events.

Pod Disruption Budgets define the minimum number of pods that must remain available during voluntary disruptions such as node drains, upgrades or scaling. This prevents maintenance operations from evicting too many replicas simultaneously, directly supporting the resilience requirement.

Circuit Breaker pattern to prevent cascading failures when downstream services are struggling

The Circuit Breaker pattern trips after repeated downstream failures, short-circuiting further calls so struggling services are not overwhelmed and failures do not cascade. This satisfies the requirement to handle transient failures gracefully across microservices on OKE.

Bulkhead pattern to isolate resource pools (e.g., thread pools) so one failing service doesn't starve others

Bulkhead partitioning allocates separate thread pools per downstream dependency, so a slow or failing service exhausts only its own pool rather than consuming all threads. This directly satisfies the stem's requirement to handle transient failures gracefully by preventing cascading resource starvation across microservices on OKE.

Retries with Exponential Backoff and Jitter for transient network or service errors

Retries with exponential backoff and jitter reattempt transient network or service errors, increasing delay between attempts and randomising timing to avoid thundering-herd synchronisation. This directly satisfies the stem's requirement to handle transient failures gracefully, since most such errors resolve on retry without manual intervention.

Use the `PARALLEL` parameter during both export and import operations to utilize multiple worker processes.

The PARALLEL parameter starts multiple worker processes that read and write Data Pump files concurrently, so export and import throughput scales with available CPU and I/O rather than running single-threaded. This directly addresses the performance-optimisation constraint for large database migrations over network connections.

Enable data compression using the `COMPRESS` parameter to minimize network transfer times.

The COMPRESS parameter shrinks dump file data before it crosses the network, so fewer bytes transfer and overall elapsed time falls. That reduction in transferred volume is the mechanism satisfying the stem's requirement to optimise export and import performance over network connections.

Use network-linked Data Pump (`NETWORK_LINK`) to stream data directly from source to target database without generating intermediate dump files on disk.

Network-linked Data Pump streams table data directly between the source and target databases over the network, bypassing the intermediate dump file stage. This removes the disk write and read overhead on the host, directly satisfying the stem's requirement to optimise export and import performance across network connections.

Using route distributions to control which routes are imported into or exported from a DRG route table.

Route distributions define which routes a DRG route table imports from an attachment or exports to it, letting architects filter and scope reachability per attachment. This is a supported DRG routing concept, distinct from static route entries added directly to the table.

Importing static or dynamic routes from FastConnect and VPN attachments into DRG route tables.

DRG route tables accept route imports from FastConnect virtual circuits and IPSec VPN attachments, learning on-premises static and dynamic (BGP) routes. This satisfies the hybrid connectivity requirement by propagating reachability into the DRG without manual route entry.

OCI Command Line Interface (CLI) and REST APIs

The OCI CLI and REST APIs expose the full Autonomous Database control surface, including provisioning, scaling and lifecycle operations, so administrators can manage instances programmatically or from a terminal. This satisfies the stem's requirement for management interfaces, complementing the OCI Console rather than replacing it.

Oracle Cloud Infrastructure (OCI) Console

The OCI Console provides a graphical web interface for provisioning, scaling, and monitoring Autonomous Database instances, satisfying the requirement for an administrative management tool. It exposes lifecycle operations such as starting, stopping, and cloning directly, so administrators can manage instances without command-line access or API scripting.

OCI Audit service to capture all control plane API activities and resource modifications.

OCI Audit service records every control plane API call and resource modification across the tenancy, providing the immutable audit trail the stem demands for security auditing and threat detection. It captures who did what, when, and from where, satisfying the comprehensive audit-trail requirement that data-plane-only logging cannot fulfil.

OCI Logging service to ingest, store, and analyze log data from VCN flow logs and compute instances.

OCI Logging ingests VCN flow logs and compute instance logs into a single, searchable service, satisfying the audit-trail and threat-detection requirement. Flow logs expose network-level anomalies such as port scanning or unexpected egress, while instance logs surface authentication and process events, giving the unified visibility the scenario demands.

Check OCI VCN DNS resolver settings and custom resolver forwarders if integrating with on-premises DNS.

Checking the VCN resolver configuration and any custom forwarders verifies that OCI-side name resolution and on-premises DNS integration are functioning. Misconfigured forwarders or resolver settings are a common cause of resolution failures during cutover, so this isolates the fault quickly.

Verify that public DNS records (A and CNAME) have been updated with the correct OCI Load Balancer IP addresses and check TTL settings.

Verifying that A and CNAME records point to the correct OCI Load Balancer IPs, and checking TTL values, confirms external name resolution reflects the migrated endpoints. Stale records or long TTLs commonly cause resolution failures immediately after cutover.

SSL/TLS certificates attached to HTTPS listeners if terminating secure traffic

HTTPS listeners require an attached SSL/TLS certificate to terminate encrypted client traffic; without it the listener cannot present a valid certificate chain. This satisfies the secure-termination requirement for the Layer 7 load balancer's high-availability configuration.

Backend sets with defined health checks to monitor backend instance responsiveness

Health checks let the load balancer detect unresponsive backend instances and stop routing to them, which is what sustains availability when a node fails. Without a backend set defining these checks, the listener has no target pool, so failover cannot occur.

Listeners configured with appropriate port and protocol (e.g., HTTP or HTTPS)

Listeners bind the load balancer's public-facing port and protocol, so HTTP or HTTPS traffic reaches the backend set. This satisfies the Layer 7 entry-point requirement: without a correctly configured listener, client requests never reach the load balancer at all.

Target (e.g., OCI Object Storage bucket)

A target defines the destination the connector delivers to, satisfying the requirement to move logs into OCI Object Storage. Service Connector Hub needs an explicit target resource — here a bucket — so it knows where to write the collected log objects. Without a target, no delivery occurs.

Source (e.g., OCI Logging service)

Service Connector Hub requires a source as one of its three mandatory components, and OCI Logging satisfies this by supplying the log data to be moved. Without a defined source, no logs enter the connector's pipeline, so the Object Storage target receives nothing. This directly meets the stem's requirement to define the origin of the transferred logs.

IAM policies granting the service connector permission to read from source and write to target

IAM policies authorise the service connector's principal to read from the source and write to the target. Without these grants, the connector cannot move logs from OCI Logging into the Object Storage bucket, so policies are a required component.

Trigger responder recipes to remediate detected security risks automatically or manually.

Cloud Guard responder recipes define automated or manual remediation actions, such as stopping a public bucket or disabling an overly permissive user, triggered when a detector recipe flags a problem. This directly improves security posture by closing detected risks.

Monitor tenancy resources for security misconfigurations and policy violations.

Cloud Guard continuously assesses tenancy resources against security zones and detector recipes, surfacing misconfigurations and policy violations as problems. This satisfies the stem's monitoring requirement by providing ongoing posture visibility across the tenancy, rather than one-off scanning, enabling remediation before exploitation.

Calculate an overall security score for the tenancy based on detected problems.

Cloud Guard aggregates detected problems across the tenancy and computes a security score, giving a quantified posture metric. This satisfies the requirement to improve visibility of overall cloud security posture by surfacing an actionable tenancy-wide score derived from problem severity and count.

Leveraging Transportable Tablespaces (TTS) combined with RMAN cross-platform conversion

Transportable Tablespaces with RMAN cross-platform conversion moves datafiles at the physical level, transferring the bulk 100 TB quickly while only metadata is exported and imported. This minimises the cutover window, satisfying the minimal-downtime constraint for the large migration.

Utilizing Oracle Zero Downtime Migration (ZDM) physical migration workflows

ZDM physical migration workflows use Data Guard or backup-based physical transfer to replicate the full 100 TB database to Exadata Database Service, then perform a fast switchover. This keeps downtime to the final cutover, satisfying the minimal-downtime requirement.

Configuring Oracle Data Guard or GoldenGate for ongoing delta synchronization prior to cutover

Data Guard or GoldenGate maintains continuous delta synchronisation from on-premises to Exadata Database Service while the bulk data copy runs, so only the final delta needs applying at cutover. This directly satisfies the minimal-downtime constraint for migrations exceeding 100 TB.

Enforce defined tags for resource creation using IAM policy conditions or tag defaults.

IAM policy conditions and tag defaults enforce defined tags at resource creation, making cost allocation reliable across compartments and preventing untagged spend from obscuring governance reporting. This underpins accurate cost observability by guaranteeing consistent tagging metadata.

Analyze spending patterns and spot cost anomalies using OCI Cost Analysis dashboards.

Cost Analysis dashboards visualise spend by compartment, service and tag, letting architects detect anomalies and trends that static reports miss. This satisfies the observability requirement by turning raw usage data into actionable insight, forming the analytical layer of the governance strategy.

Configure OCI Budgets with both actual and forecasted spending alert thresholds.

OCI Budgets track actual and forecasted spend against a defined limit, triggering alerts when either threshold is breached. This satisfies the stem's cost governance requirement by providing early warning of overspend before the billing period closes, enabling corrective action rather than retrospective reporting.

OCI Load Balancers and Network Load Balancers

Load Balancers and Network Load Balancers are supported members of an FSDR Protection Group, so their configuration and traffic steering can be orchestrated during a switchover or failover. They are among the resource types the DR plan can manage.

Oracle Database systems (Base Database and Exadata Database Service)

FSDR DR Protection Groups orchestrate Oracle Database systems, covering Base Database and Exadata Database Service, alongside compute and networking. This membership satisfies the stem's requirement for manageable resource types, since FSDR plans switchover and failover for these database systems across regions.

OCI Compute instances (including boot and block volumes)

FSDR DR Protection Groups include OCI Compute instances together with their attached boot and block volumes, so the group captures both the instance and its storage. This satisfies the stem's requirement by letting FSDR orchestrate compute failover while keeping volume data consistent.