Practice Test 3 | Google Cloud Certified Professional Cloud Architect | Dumps | Mock Test
For this question, refer to the EHR Healthcare case study.
EHR Healthcare has decided to create and maintain a monitoring system of all the services offered, wishing to provide extremely good levels of service, such as a minimum availability of 99.9% for all customer-facing systems.
To achieve this, it is setting up extremely efficient services managed with Kubernetes, and is implementing hot / warm Disaster Recovery procedures.
According to your expertise, are there any additional activities and procedures that must be activated to really ensure all this (pick 2)?
A. There are no further activities; GCP provides Dashboard to get an overview of the system
B. Create health checks on the Load Balancer VM
C. There are no further activities; when selecting services with a certain level of service, this can be offered directly to customers
D. Identify SLIs (service level indicators) and design SLOs (service level objectives) and error budget
E. Set Alert with Monitoring on all the metrics of interest, so that you can intervene in the event of an overrun
F. Use the Cloud Monitor API in Operation Suite to create increasingly serious alerts based on the SLO burn rate. Check the history of metrics and alerts in service dashboards
G. Create an additional K8s cluster on AWS for increased availability
Correct answers: D and F
SRE is the accredited methodology to keep the declared or agreed service levels under control; it involves identifying indicators (response times, latency, availability, etc.) called SLI and establishing objective SLO levels (lower than SLAs) and a margin error allowed.
Based on these values, it is therefore necessary to activate the alerts and prepare corrective actions before reaching the set limits.
- A and B are wrong because GCP offers the tools but the customer has to manage its service objectives
- B is wrong because it may be only a small part of the overall strategy.
- C is wrong because the service level of the GCP services cannot be transferred directly to the customer applications, that may have its own faults
- E is not completely exact, because the alert of an overrun may not be enough
- G is wrong because it is a generic action not supported by metrics and methodology
For any further detail:
- https://cloud.google.com/stackdriver/docs/solutions/slo-monitoring/sli-metrics/overview
- https://cloud.google.com/stackdriver/docs/solutions/slo-monitoring
- https://cloud.google.com/blog/products/devops-sre/sre-fundamentals-sli-vs-slo-vs-sla
- https://cloud.google.com/stackdriver/docs/solutions/slo-monitoring
- https://cloud.google.com/architecture/devops/devops-measurement-proactive-failure-notification
Comments are closed, but trackbacks and pingbacks are open.