cert
cert-1
cert-2

Pass Cisco 300-640 Exam in First Attempt Guaranteed!

Get 100% Latest Exam Questions, Accurate & Verified Answers to Pass the Actual Exam!
30 Days Free Updates, Instant Download!

cert-5
cert-6
300-640 Exam - Verified By Experts
300-640 Premium File

300-640 Premium File

$89.99
$98.99
  • Premium File 60 Questions & Answers. Last Update: Nov 10, 2026

Whats Included:

  • Latest Questions
  • 100% Accurate Answers
  • Fast Exam Updates
 
$98.99
$89.99
accept 10 downloads in the last 7 days
block-screenshots
Cisco 300-640 Practice Test 59

Question 580

The acceptance board must decide whether an AI-infrastructure response to cross-system clock drift diagnosis is fit for controlled operation. The fictional operating facts are: A new AI cluster has intermittent signed API authentication failures and its GPU nodes disagree with fabric management time. Signoff cannot rely on a good-looking utilization graph, a change ticket alone or an unmeasured promise. The response must include an appropriate governing engineering principle, a scoped intervention and a separate measurable acceptance criterion. Choose EXACTLY THREE statements that meet those roles.

  1. Restore supported redundant time synchronization and verify event ordering before further root-cause changes.
  2. Sustained training no longer throttles and rack electrical limits remain satisfied.
  3. Cross-system logs align within tolerance and scheduled job authentication succeeds.
  4. Operational telemetry and certificate validation depend on reliable time across fabric and compute components.
  5. Reliable troubleshooting requires a synchronized incident timeline spanning application, GPU, network, compute and storage telemetry.
  6. Stop rollout on breached acceptance criteria and restore a tested configuration before root-cause rework.

Explanation

Correct Answer

A:

A new AI cluster has intermittent signed API authentication failures and its GPU nodes disagree with fabric management time. Restore supported redundant time synchronization and verify event ordering before further root-cause changes. Operational telemetry and certificate validation depend on reliable time across fabric and compute components.

C:

A new AI cluster has intermittent signed API authentication failures and its GPU nodes disagree with fabric management time. Cross-system logs align within tolerance and scheduled job authentication succeeds. Restore supported redundant time synchronization and verify event ordering before further root-cause changes.

D:

A new AI cluster has intermittent signed API authentication failures and its GPU nodes disagree with fabric management time. Operational telemetry and certificate validation depend on reliable time across fabric and compute components. Restore supported redundant time synchronization and verify event ordering before further root-cause changes.

Incorrect Answers

B:

Sustained training no longer throttles and rack electrical limits remain satisfied. It would not independently demonstrate that the control works: A new AI cluster has intermittent signed API authentication failures and its GPU nodes disagree with fabric management time. Cross-system logs align within tolerance and scheduled job authentication succeeds. Restore supported redundant time synchronization and verify event ordering before further root-cause changes.

E:

Reliable troubleshooting requires a synchronized incident timeline spanning application, GPU, network, compute and storage telemetry. It is not the rule that supports the proposed change and acceptance test: A new AI cluster has intermittent signed API authentication failures and its GPU nodes disagree with fabric management time. Operational telemetry and certificate validation depend on reliable time across fabric and compute components. Restore supported redundant time synchronization and verify event ordering before further root-cause changes.

F:

Stop rollout on breached acceptance criteria and restore a tested configuration before root-cause rework. It does not implement the governing rule for this approval: A new AI cluster has intermittent signed API authentication failures and its GPU nodes disagree with fabric management time. Restore supported redundant time synchronization and verify event ordering before further root-cause changes. Cross-system logs align within tolerance and scheduled job authentication succeeds.

Question 581

A data-center AI architecture committee considers thermal and power event correlation in its AI Infrastructure Operations and Troubleshooting review. At the affected site: A training cluster's throughput drops each afternoon as warm intake air drives accelerator clock throttling. The committee is deliberately not selecting a vendor product or applying a live change yet. It needs the foundational reasoning that makes the next design decision defensible. Which ONE engineering principle directly governs this situation? Choose ONE answer.

  1. Operational telemetry and certificate validation depend on reliable time across fabric and compute components.
  2. Operational changes should have independent service acceptance and a bounded rollback trigger rather than trust a successful automation response.
  3. Sporadic AI slowdown can result from thermal throttling or electrical power caps rather than code regression.
  4. Network operations should detect degrading AI fabric conditions using correlated interface, queue and topology signals.

Explanation

Correct Answer

C:

A training cluster's throughput drops each afternoon as warm intake air drives accelerator clock throttling. Sporadic AI slowdown can result from thermal throttling or electrical power caps rather than code regression. Correlate GPU clock rate, power limit flags, inlet temperature, UCS faults and workload intensity.

Incorrect Answers

A:

Operational telemetry and certificate validation depend on reliable time across fabric and compute components. It does not govern the condition described: A training cluster's throughput drops each afternoon as warm intake air drives accelerator clock throttling. Sporadic AI slowdown can result from thermal throttling or electrical power caps rather than code regression.

B:

Operational changes should have independent service acceptance and a bounded rollback trigger rather than trust a successful automation response. It does not govern the condition described: A training cluster's throughput drops each afternoon as warm intake air drives accelerator clock throttling. Sporadic AI slowdown can result from thermal throttling or electrical power caps rather than code regression.

D:

Network operations should detect degrading AI fabric conditions using correlated interface, queue and topology signals. It does not govern the condition described: A training cluster's throughput drops each afternoon as warm intake air drives accelerator clock throttling. Sporadic AI slowdown can result from thermal throttling or electrical power caps rather than code regression.

Question 582

An infrastructure diagnostician is investigating thermal and power event correlation for the AI Infrastructure Operations and Troubleshooting team. The reported incident reads: A training cluster's throughput drops each afternoon as warm intake air drives accelerator clock throttling. Competing teams attribute the symptoms to different layers of the AI stack. Before accepting a hypothesis or touching production settings, the investigator must choose a discriminating set of measurements. Which ONE evidence package is MOST relevant? Choose ONE answer.

  1. Trace per-priority pause duration, ingress queue depth, flow source and blocked unrelated classes across hops.
  2. Compare known-good and canary job metrics, error rates, fabric health and rollback completion time.
  3. Correlate GPU clock rate, power limit flags, inlet temperature, UCS faults and workload intensity.
  4. Inspect Intersight equipment faults, fan or PSU telemetry, NIC error counters, component history and service-profile association.

Explanation

Correct Answer

C:

A training cluster's throughput drops each afternoon as warm intake air drives accelerator clock throttling. Correlate GPU clock rate, power limit flags, inlet temperature, UCS faults and workload intensity. Sporadic AI slowdown can result from thermal throttling or electrical power caps rather than code regression.

Incorrect Answers

A:

Trace per-priority pause duration, ingress queue depth, flow source and blocked unrelated classes across hops. That check would not establish the decisive facts for this condition: A training cluster's throughput drops each afternoon as warm intake air drives accelerator clock throttling. Correlate GPU clock rate, power limit flags, inlet temperature, UCS faults and workload intensity.

B:

Compare known-good and canary job metrics, error rates, fabric health and rollback completion time. That check would not establish the decisive facts for this condition: A training cluster's throughput drops each afternoon as warm intake air drives accelerator clock throttling. Correlate GPU clock rate, power limit flags, inlet temperature, UCS faults and workload intensity.

D:

Inspect Intersight equipment faults, fan or PSU telemetry, NIC error counters, component history and service-profile association. That check would not establish the decisive facts for this condition: A training cluster's throughput drops each afternoon as warm intake air drives accelerator clock throttling. Correlate GPU clock rate, power limit flags, inlet temperature, UCS faults and workload intensity.

Question 583

During a bounded AI-platform change-control review concerning thermal and power event correlation, the site owner reports: A training cluster's throughput drops each afternoon as warm intake air drives accelerator clock throttling. The proposed work must have a narrow implementation scope, first be validated against the failure evidence and avoid unrelated system-wide tuning. Which ONE intervention should the engineers pilot before any wider rollout? Choose ONE answer.

  1. Restore validated cooling or power headroom after facility review and verify under sustained load.
  2. Repair the measured inconsistency on the affected path and retest with the same RDMA workload.
  3. Adjust the proven bottleneck and validated marking behavior instead of raising thresholds blindly.
  4. Normalize clocks and correlate the first causal infrastructure event with impacted job transactions.

Explanation

Correct Answer

A:

A training cluster's throughput drops each afternoon as warm intake air drives accelerator clock throttling. Restore validated cooling or power headroom after facility review and verify under sustained load. Correlate GPU clock rate, power limit flags, inlet temperature, UCS faults and workload intensity.

Incorrect Answers

B:

Repair the measured inconsistency on the affected path and retest with the same RDMA workload. That change would leave the reported problem unresolved: A training cluster's throughput drops each afternoon as warm intake air drives accelerator clock throttling. Restore validated cooling or power headroom after facility review and verify under sustained load.

C:

Adjust the proven bottleneck and validated marking behavior instead of raising thresholds blindly. That change would leave the reported problem unresolved: A training cluster's throughput drops each afternoon as warm intake air drives accelerator clock throttling. Restore validated cooling or power headroom after facility review and verify under sustained load.

D:

Normalize clocks and correlate the first causal infrastructure event with impacted job transactions. That change would leave the reported problem unresolved: A training cluster's throughput drops each afternoon as warm intake air drives accelerator clock throttling. Restore validated cooling or power headroom after facility review and verify under sustained load.

Question 584

Two independent technical reviewers examine a AI Infrastructure Operations and Troubleshooting problem involving thermal and power event correlation. Their shared incident log states: A training cluster's throughput drops each afternoon as warm intake air drives accelerator clock throttling. One reviewer requires measurements capable of separating the observed problem from neighboring failure modes; the other requires a bounded action that follows those measurements. Choose EXACTLY TWO statements that together supply those two parts.

  1. Repair the measured inconsistency on the affected path and retest with the same RDMA workload.
  2. Check shared correlation IDs, NTP offsets, job start times, network event sequence and storage logs.
  3. Correlate GPU clock rate, power limit flags, inlet temperature, UCS faults and workload intensity.
  4. Restore validated cooling or power headroom after facility review and verify under sustained load.
  5. Measure NTP reachability, UTC offset, source hierarchy, log timestamps and drift during node restarts.

Explanation

Correct Answer

C:

A training cluster's throughput drops each afternoon as warm intake air drives accelerator clock throttling. Correlate GPU clock rate, power limit flags, inlet temperature, UCS faults and workload intensity. Sustained training no longer throttles and rack electrical limits remain satisfied.

D:

A training cluster's throughput drops each afternoon as warm intake air drives accelerator clock throttling. Restore validated cooling or power headroom after facility review and verify under sustained load. Sustained training no longer throttles and rack electrical limits remain satisfied.

Incorrect Answers

A:

Repair the measured inconsistency on the affected path and retest with the same RDMA workload. It is not the scoped response supported by the accompanying evidence: A training cluster's throughput drops each afternoon as warm intake air drives accelerator clock throttling. Restore validated cooling or power headroom after facility review and verify under sustained load. Sustained training no longer throttles and rack electrical limits remain satisfied.

B:

Check shared correlation IDs, NTP offsets, job start times, network event sequence and storage logs. It would not provide the evidence needed for the paired diagnosis and follow-up: A training cluster's throughput drops each afternoon as warm intake air drives accelerator clock throttling. Correlate GPU clock rate, power limit flags, inlet temperature, UCS faults and workload intensity. Sustained training no longer throttles and rack electrical limits remain satisfied.

E:

Measure NTP reachability, UTC offset, source hierarchy, log timestamps and drift during node restarts. It would not provide the evidence needed for the paired diagnosis and follow-up: A training cluster's throughput drops each afternoon as warm intake air drives accelerator clock throttling. Correlate GPU clock rate, power limit flags, inlet temperature, UCS faults and workload intensity. Sustained training no longer throttles and rack electrical limits remain satisfied.

Question 585

The acceptance board must decide whether an AI-infrastructure response to thermal and power event correlation is fit for controlled operation. The fictional operating facts are: A training cluster's throughput drops each afternoon as warm intake air drives accelerator clock throttling. Signoff cannot rely on a good-looking utilization graph, a change ticket alone or an unmeasured promise. The response must include an appropriate governing engineering principle, a scoped intervention and a separate measurable acceptance criterion. Choose EXACTLY THREE statements that meet those roles.

  1. Repeated trials demonstrate a statistically credible throughput or latency effect within operational guardrails.
  2. RDMA performance faults require mapping NIC, switch and remote endpoint telemetry along the actual routed path.
  3. Sporadic AI slowdown can result from thermal throttling or electrical power caps rather than code regression.
  4. Restore validated cooling or power headroom after facility review and verify under sustained load.
  5. Sustained training no longer throttles and rack electrical limits remain satisfied.
  6. Run representative controlled benchmarks before and after the change with documented normalization.

Explanation

Correct Answer

C:

A training cluster's throughput drops each afternoon as warm intake air drives accelerator clock throttling. Sporadic AI slowdown can result from thermal throttling or electrical power caps rather than code regression. Restore validated cooling or power headroom after facility review and verify under sustained load.

D:

A training cluster's throughput drops each afternoon as warm intake air drives accelerator clock throttling. Restore validated cooling or power headroom after facility review and verify under sustained load. Sporadic AI slowdown can result from thermal throttling or electrical power caps rather than code regression.

E:

A training cluster's throughput drops each afternoon as warm intake air drives accelerator clock throttling. Sustained training no longer throttles and rack electrical limits remain satisfied. Restore validated cooling or power headroom after facility review and verify under sustained load.

Incorrect Answers

A:

Repeated trials demonstrate a statistically credible throughput or latency effect within operational guardrails. It would not independently demonstrate that the control works: A training cluster's throughput drops each afternoon as warm intake air drives accelerator clock throttling. Sustained training no longer throttles and rack electrical limits remain satisfied. Restore validated cooling or power headroom after facility review and verify under sustained load.

B:

RDMA performance faults require mapping NIC, switch and remote endpoint telemetry along the actual routed path. It is not the rule that supports the proposed change and acceptance test: A training cluster's throughput drops each afternoon as warm intake air drives accelerator clock throttling. Sporadic AI slowdown can result from thermal throttling or electrical power caps rather than code regression. Restore validated cooling or power headroom after facility review and verify under sustained load.

F:

Run representative controlled benchmarks before and after the change with documented normalization. It does not implement the governing rule for this approval: A training cluster's throughput drops each afternoon as warm intake air drives accelerator clock throttling. Restore validated cooling or power headroom after facility review and verify under sustained load. Sustained training no longer throttles and rack electrical limits remain satisfied.

Question 586

A data-center AI architecture committee considers thermal and power event correlation in its AI Infrastructure Operations and Troubleshooting review. At the affected site: A model-serving farm slows after a new power cap is applied during rack consolidation but application metrics show no code change. The committee is deliberately not selecting a vendor product or applying a live change yet. It needs the foundational reasoning that makes the next design decision defensible. Which ONE engineering principle directly governs this situation? Choose ONE answer.

  1. GPU idle periods may originate in storage or input pipelines, so remediation needs cross-layer timing evidence.
  2. Sporadic AI slowdown can result from thermal throttling or electrical power caps rather than code regression.
  3. Server health incidents need correlated power, thermal, adapter and inventory evidence before blaming the AI model.
  4. Frequent priority pauses can indicate congestion spreading or head-of-line blocking rather than healthy lossless delivery.

Explanation

Correct Answer

B:

A model-serving farm slows after a new power cap is applied during rack consolidation but application metrics show no code change. Sporadic AI slowdown can result from thermal throttling or electrical power caps rather than code regression. Correlate GPU clock rate, power limit flags, inlet temperature, UCS faults and workload intensity.

Incorrect Answers

A:

GPU idle periods may originate in storage or input pipelines, so remediation needs cross-layer timing evidence. It does not govern the condition described: A model-serving farm slows after a new power cap is applied during rack consolidation but application metrics show no code change. Sporadic AI slowdown can result from thermal throttling or electrical power caps rather than code regression.

C:

Server health incidents need correlated power, thermal, adapter and inventory evidence before blaming the AI model. It does not govern the condition described: A model-serving farm slows after a new power cap is applied during rack consolidation but application metrics show no code change. Sporadic AI slowdown can result from thermal throttling or electrical power caps rather than code regression.

D:

Frequent priority pauses can indicate congestion spreading or head-of-line blocking rather than healthy lossless delivery. It does not govern the condition described: A model-serving farm slows after a new power cap is applied during rack consolidation but application metrics show no code change. Sporadic AI slowdown can result from thermal throttling or electrical power caps rather than code regression.

Question 587

An infrastructure diagnostician is investigating thermal and power event correlation for the AI Infrastructure Operations and Troubleshooting team. The reported incident reads: A model-serving farm slows after a new power cap is applied during rack consolidation but application metrics show no code change. Competing teams attribute the symptoms to different layers of the AI stack. Before accepting a hypothesis or touching production settings, the investigator must choose a discriminating set of measurements. Which ONE evidence package is MOST relevant? Choose ONE answer.

  1. Correlate GPU clock rate, power limit flags, inlet temperature, UCS faults and workload intensity.
  2. Compare known-good and canary job metrics, error rates, fabric health and rollback completion time.
  3. Plot switch queue depth, ECN marks, packet drops, RTT and endpoint rate response over synchronized intervals.
  4. Compare data-loader wait, storage read p99, metadata latency, network path and GPU kernel occupancy.

Explanation

Correct Answer

A:

A model-serving farm slows after a new power cap is applied during rack consolidation but application metrics show no code change. Correlate GPU clock rate, power limit flags, inlet temperature, UCS faults and workload intensity. Sporadic AI slowdown can result from thermal throttling or electrical power caps rather than code regression.

Incorrect Answers

B:

Compare known-good and canary job metrics, error rates, fabric health and rollback completion time. That check would not establish the decisive facts for this condition: A model-serving farm slows after a new power cap is applied during rack consolidation but application metrics show no code change. Correlate GPU clock rate, power limit flags, inlet temperature, UCS faults and workload intensity.

C:

Plot switch queue depth, ECN marks, packet drops, RTT and endpoint rate response over synchronized intervals. That check would not establish the decisive facts for this condition: A model-serving farm slows after a new power cap is applied during rack consolidation but application metrics show no code change. Correlate GPU clock rate, power limit flags, inlet temperature, UCS faults and workload intensity.

D:

Compare data-loader wait, storage read p99, metadata latency, network path and GPU kernel occupancy. That check would not establish the decisive facts for this condition: A model-serving farm slows after a new power cap is applied during rack consolidation but application metrics show no code change. Correlate GPU clock rate, power limit flags, inlet temperature, UCS faults and workload intensity.

Question 588

During a bounded AI-platform change-control review concerning thermal and power event correlation, the site owner reports: A model-serving farm slows after a new power cap is applied during rack consolidation but application metrics show no code change. The proposed work must have a narrow implementation scope, first be validated against the failure evidence and avoid unrelated system-wide tuning. Which ONE intervention should the engineers pilot before any wider rollout? Choose ONE answer.

  1. Tune symptom-based alerts and route to responsible teams with runbook-linked evidence.
  2. Normalize clocks and correlate the first causal infrastructure event with impacted job transactions.
  3. Restore validated cooling or power headroom after facility review and verify under sustained load.
  4. Run representative controlled benchmarks before and after the change with documented normalization.

Explanation

Correct Answer

C:

A model-serving farm slows after a new power cap is applied during rack consolidation but application metrics show no code change. Restore validated cooling or power headroom after facility review and verify under sustained load. Correlate GPU clock rate, power limit flags, inlet temperature, UCS faults and workload intensity.

Incorrect Answers

A:

Tune symptom-based alerts and route to responsible teams with runbook-linked evidence. That change would leave the reported problem unresolved: A model-serving farm slows after a new power cap is applied during rack consolidation but application metrics show no code change. Restore validated cooling or power headroom after facility review and verify under sustained load.

B:

Normalize clocks and correlate the first causal infrastructure event with impacted job transactions. That change would leave the reported problem unresolved: A model-serving farm slows after a new power cap is applied during rack consolidation but application metrics show no code change. Restore validated cooling or power headroom after facility review and verify under sustained load.

D:

Run representative controlled benchmarks before and after the change with documented normalization. That change would leave the reported problem unresolved: A model-serving farm slows after a new power cap is applied during rack consolidation but application metrics show no code change. Restore validated cooling or power headroom after facility review and verify under sustained load.

Question 589

Two independent technical reviewers examine a AI Infrastructure Operations and Troubleshooting problem involving thermal and power event correlation. Their shared incident log states: A model-serving farm slows after a new power cap is applied during rack consolidation but application metrics show no code change. One reviewer requires measurements capable of separating the observed problem from neighboring failure modes; the other requires a bounded action that follows those measurements. Choose EXACTLY TWO statements that together supply those two parts.

  1. Correlate GPU clock rate, power limit flags, inlet temperature, UCS faults and workload intensity.
  2. Restore validated cooling or power headroom after facility review and verify under sustained load.
  3. Repair the measured inconsistency on the affected path and retest with the same RDMA workload.
  4. Plot switch queue depth, ECN marks, packet drops, RTT and endpoint rate response over synchronized intervals.
  5. Compare data-loader wait, storage read p99, metadata latency, network path and GPU kernel occupancy.

Explanation

Correct Answer

A:

A model-serving farm slows after a new power cap is applied during rack consolidation but application metrics show no code change. Correlate GPU clock rate, power limit flags, inlet temperature, UCS faults and workload intensity. Sustained training no longer throttles and rack electrical limits remain satisfied.

B:

A model-serving farm slows after a new power cap is applied during rack consolidation but application metrics show no code change. Restore validated cooling or power headroom after facility review and verify under sustained load. Sustained training no longer throttles and rack electrical limits remain satisfied.

Incorrect Answers

C:

Repair the measured inconsistency on the affected path and retest with the same RDMA workload. It is not the scoped response supported by the accompanying evidence: A model-serving farm slows after a new power cap is applied during rack consolidation but application metrics show no code change. Restore validated cooling or power headroom after facility review and verify under sustained load. Sustained training no longer throttles and rack electrical limits remain satisfied.

D:

Plot switch queue depth, ECN marks, packet drops, RTT and endpoint rate response over synchronized intervals. It would not provide the evidence needed for the paired diagnosis and follow-up: A model-serving farm slows after a new power cap is applied during rack consolidation but application metrics show no code change. Correlate GPU clock rate, power limit flags, inlet temperature, UCS faults and workload intensity. Sustained training no longer throttles and rack electrical limits remain satisfied.

E:

Compare data-loader wait, storage read p99, metadata latency, network path and GPU kernel occupancy. It would not provide the evidence needed for the paired diagnosis and follow-up: A model-serving farm slows after a new power cap is applied during rack consolidation but application metrics show no code change. Correlate GPU clock rate, power limit flags, inlet temperature, UCS faults and workload intensity. Sustained training no longer throttles and rack electrical limits remain satisfied.

Question 590

The acceptance board must decide whether an AI-infrastructure response to thermal and power event correlation is fit for controlled operation. The fictional operating facts are: A model-serving farm slows after a new power cap is applied during rack consolidation but application metrics show no code change. Signoff cannot rely on a good-looking utilization graph, a change ticket alone or an unmeasured promise. The response must include an appropriate governing engineering principle, a scoped intervention and a separate measurable acceptance criterion. Choose EXACTLY THREE statements that meet those roles.

  1. Sporadic AI slowdown can result from thermal throttling or electrical power caps rather than code regression.
  2. Sustained training no longer throttles and rack electrical limits remain satisfied.
  3. Restore validated cooling or power headroom after facility review and verify under sustained load.
  4. The repaired node rejoins service with normal health and controlled workload restoration.
  5. Correct the evidenced storage or pipeline bottleneck before increasing accelerator capacity.
  6. Congestion remediation must distinguish healthy ECN signaling from persistent oversubscription and unstable thresholds.

Explanation

Correct Answer

A:

A model-serving farm slows after a new power cap is applied during rack consolidation but application metrics show no code change. Sporadic AI slowdown can result from thermal throttling or electrical power caps rather than code regression. Restore validated cooling or power headroom after facility review and verify under sustained load.

B:

A model-serving farm slows after a new power cap is applied during rack consolidation but application metrics show no code change. Sustained training no longer throttles and rack electrical limits remain satisfied. Restore validated cooling or power headroom after facility review and verify under sustained load.

C:

A model-serving farm slows after a new power cap is applied during rack consolidation but application metrics show no code change. Restore validated cooling or power headroom after facility review and verify under sustained load. Sporadic AI slowdown can result from thermal throttling or electrical power caps rather than code regression.

Incorrect Answers

D:

The repaired node rejoins service with normal health and controlled workload restoration. It would not independently demonstrate that the control works: A model-serving farm slows after a new power cap is applied during rack consolidation but application metrics show no code change. Sustained training no longer throttles and rack electrical limits remain satisfied. Restore validated cooling or power headroom after facility review and verify under sustained load.

E:

Correct the evidenced storage or pipeline bottleneck before increasing accelerator capacity. It does not implement the governing rule for this approval: A model-serving farm slows after a new power cap is applied during rack consolidation but application metrics show no code change. Restore validated cooling or power headroom after facility review and verify under sustained load. Sustained training no longer throttles and rack electrical limits remain satisfied.

F:

Congestion remediation must distinguish healthy ECN signaling from persistent oversubscription and unstable thresholds. It is not the rule that supports the proposed change and acceptance test: A model-serving farm slows after a new power cap is applied during rack consolidation but application metrics show no code change. Sporadic AI slowdown can result from thermal throttling or electrical power caps rather than code regression. Restore validated cooling or power headroom after facility review and verify under sustained load.

Implementing Cisco Data Center AI Infrastructure (DCAI) (300-640) Practice Tests