Question 560
A reliability report must include an observed numerical success metric and a quantified target for that metric. Which TWO service-level concepts correspond to those roles? Choose TWO.
- Alert thresholds relative to the SLO and SLI
- Service-level agreement (SLA)
- Service-level objective (SLO)
- Allowable error budget
- Service-level indicator (SLI)
Explanation
Correct Answers
C:An SLO specifies the desired target for a service-level indicator over a defined period, such as 99.95 percent successful requests. The SLI provides the observed measurements that can be compared with that target.
E:An SLI is a quantitative measure of actual service behavior, such as success ratio or latency. It is the signal compared against a target.
Incorrect Answers
A:Monitoring thresholds must reflect the reliability target and measurement window; the requirement is pairing observed service quality with a reliability target, outside the effect of this choice.
B:An SLA is a formal commitment between provider and customer with terms, responsibilities and potential remedies; it does not deliver pairing observed service quality with a reliability target, which is decisive in this scenario.
D:An error budget represents the tolerated time or request failures implied by an SLO; the task involves pairing observed service quality with a reliability target, whereas this addresses another subsystem.
Question 561
An operations agreement specifies contractual availability remedies if the vendor falls below a promised level. Which term describes the agreement?
- Service-level indicator (SLI)
- Service-level agreement (SLA)
- Successful-request ratio
- Allowable error budget
Explanation
Correct Answer
B:An SLA is a formal commitment between provider and customer with terms, responsibilities and potential remedies. It is distinct from an internal operational target.
Incorrect Answers
A:An SLI is a quantitative measure of actual service behavior, such as success ratio or latency; that is not equivalent to identifying a contractual service commitment.
C:Counting application-level successful responses detects failures even if the process is running; it answers a neighboring question rather than identifying a contractual service commitment.
D:An error budget represents the tolerated time or request failures implied by an SLO; that changes a different part of the system; the requested outcome is identifying a contractual service commitment.
Question 562
A service was deployed just before errors spiked, and responders want to link the time of change to the graph. What data improves analysis?
- Version-controlled monitoring configuration
- Sustained-duration alert evaluation
- Telemetry freshness or missing-data alarm
- Deployment event annotations
Explanation
Correct Answer
D:Change events overlaid on telemetry provide context for correlation. Time coincidence still requires investigation before assigning causality.
Incorrect Answers
A:Tracking alert rules and their review history supports rollback and reproducibility; it will not provide correlating metric changes with deployments under these conditions.
B:Requiring a condition to remain true for a specified interval filters transient bursts; using it would leave correlating metric changes with deployments unresolved.
C:Monitoring the absence of expected reports detects failures in the agent or data pipeline; the requirement is correlating metric changes with deployments, outside the effect of this choice.
Question 563
A device should send an unsolicited management notification when an interface fails. Which SNMP mechanism supports this?
- SNMP polling
- Central log aggregation
- SNMP management information bases (MIBs)
- SNMP trap
Explanation
Correct Answer
D:A trap sends an event notification from an agent to a manager. Traps may be lost, so monitoring often combines them with polling.
Incorrect Answers
A:SNMP can retrieve indexed management variables, such as traffic counters, from managed devices; the task involves delivering pushed network-device events, whereas this addresses another subsystem.
B:Aggregation brings events together with timestamps and source identities to support correlation; its scope differs from delivering pushed network-device events, so it cannot meet the stated requirement.
C:MIBs describe the names, types and meanings of SNMP-managed objects; that is not equivalent to delivering pushed network-device events.
Question 564
A monitoring dashboard shows uptime but ignores failed transactions that return HTTP 500. Which indicator best closes this gap?
- Alert thresholds relative to the SLO and SLI
- Successful-request ratio
- Service-level indicator (SLI)
- Allowable error budget
- Service-level agreement (SLA)
Explanation
Correct Answer
B:Counting application-level successful responses detects failures even if the process is running. Liveness alone does not guarantee correct service behavior.
Incorrect Answers
A:Monitoring thresholds must reflect the reliability target and measurement window; that changes a different part of the system; the requested outcome is measuring application outcomes rather than process presence.
C:An SLI is a quantitative measure of actual service behavior, such as success ratio or latency; it will not provide measuring application outcomes rather than process presence under these conditions.
D:An error budget represents the tolerated time or request failures implied by an SLO; using it would leave measuring application outcomes rather than process presence unresolved.
E:An SLA is a formal commitment between provider and customer with terms, responsibilities and potential remedies; the requirement is measuring application outcomes rather than process presence, outside the effect of this choice.
Question 565
A monitoring system receives an HTTPS callback when a deployment completes. Which delivery pattern is being used?
- Application health check
- SNMP management information bases (MIBs)
- Webhook notification
- SNMP trap
Explanation
Correct Answer
C:A webhook sends an HTTP request in response to an event. It reduces polling latency but the receiver needs authentication and retry handling.
Incorrect Answers
A:Health checks test a service's declared readiness or liveness with explicit success criteria; it does not deliver receiving event-driven HTTP callbacks, which is decisive in this scenario.
B:MIBs describe the names, types and meanings of SNMP-managed objects; the evidence calls for receiving event-driven HTTP callbacks, not merely this separate operation.
D:A trap sends an event notification from an agent to a manager; its scope differs from receiving event-driven HTTP callbacks, so it cannot meet the stated requirement.
Question 566
A team must interpret vendor-specific SNMP object identifiers using named schemas. Which documents define the object structure?
- Central log aggregation
- SNMP management information bases (MIBs)
- Application health check
- SNMP polling
Explanation
Correct Answer
B:MIBs describe the names, types and meanings of SNMP-managed objects. They do not collect metrics by themselves.
Incorrect Answers
A:Aggregation brings events together with timestamps and source identities to support correlation; the administrator instead needs resolving SNMP identifiers to named variables; this option serves a different purpose.
C:Health checks test a service's declared readiness or liveness with explicit success criteria; that changes a different part of the system; the requested outcome is resolving SNMP identifiers to named variables.
D:SNMP can retrieve indexed management variables, such as traffic counters, from managed devices; this does not achieve resolving SNMP identifiers to named variables on the system described.
Question 567
A monitoring server needs a lightweight probe that checks whether an HTTP /health endpoint returns expected content. Which method is suitable?
- SNMP trap
- SNMP polling
- Application health check
- Webhook notification
Explanation
Correct Answer
C:Health checks test a service's declared readiness or liveness with explicit success criteria. They detect application failures that basic ICMP probes may miss.
Incorrect Answers
A:A trap sends an event notification from an agent to a manager; using it would leave checking real application health endpoints unresolved.
B:SNMP can retrieve indexed management variables, such as traffic counters, from managed devices; the requirement is checking real application health endpoints, outside the effect of this choice.
D:A webhook sends an HTTP request in response to an event; the evidence calls for checking real application health endpoints, not merely this separate operation.
Question 568
A NIC disappeared after a new kernel boot, and lsmod no longer lists its driver. Which cause is plausible?
- Missing or failed kernel driver module
- CPU, memory, or platform hardware fault
- Driver binding and /dev node permissions
- Nested virtualization exposure
Explanation
Correct Answer
A:A changed kernel or module package can leave a controller without its expected driver. Device enumeration and kernel logs distinguish this from a DNS issue.
Incorrect Answers
B:Machine-check records can report low-level hardware errors rather than application exceptions; that is not equivalent to diagnosing a network controller without its driver.
C:USB enumeration confirms discovery but not successful driver operation or user access; the administrator instead needs diagnosing a network controller without its driver; this option serves a different purpose.
D:Guests need virtualization extensions exposed by the host hypervisor to run nested accelerated VMs; it answers a neighboring question rather than diagnosing a network controller without its driver.
Question 569
A service logs input/output error on random files, and SMART monitoring shows reallocated sectors increasing. What is the strongest suspicion?
- NFS server or network path availability
- Failing storage hardware
- Container writable-layer or quota capacity
- Kernel remounted filesystem read-only after errors
Explanation
Correct Answer
B:Increasing device errors and sector remapping suggest media degradation. Protect recoverable data before extensive filesystem repairs.
Incorrect Answers
A:A hard network mount can wait on unavailable remote I/O; this does not achieve attributing random I/O errors to drive degradation on the system described.
C:A container's storage driver or configured filesystem quota can limit writes independently of other host mounts; using it would leave attributing random I/O errors to drive degradation unresolved.
D:Some filesystems remount read-only after serious consistency or storage faults; the requirement is attributing random I/O errors to drive degradation, outside the effect of this choice.
Question 570
A storage controller is visible in lspci but no /dev block node appears. Which subsystem is the next evidence source?
- Inspect BMC power and hardware sensor events
- Cooling, airflow and fan health
- CPU, memory, or platform hardware fault
- Nested virtualization exposure
- Controller driver binding and dmesg errors
Explanation
Correct Answer
E:PCI enumeration proves the hardware is discovered, not that a driver has initialized attached storage. Kernel messages reveal binding and device-probe failures.
Incorrect Answers
A:Out-of-band management can report chassis power, thermal and hardware faults even when the OS is unavailable; the evidence calls for explaining a PCI controller with no disk nodes, not merely this separate operation.
B:Thermal throttling reduces frequency to protect silicon; the task involves explaining a PCI controller with no disk nodes, whereas this addresses another subsystem.
C:Machine-check records can report low-level hardware errors rather than application exceptions; its scope differs from explaining a PCI controller with no disk nodes, so it cannot meet the stated requirement.
D:Guests need virtualization extensions exposed by the host hypervisor to run nested accelerated VMs; that is not equivalent to explaining a PCI controller with no disk nodes.