AMAADOR ACADEMYCertificatesVerify a certificate

SRECD

Final check — SRE, Observability and CI/CD Pipelines

Answer the 40 questions. You need 80% or more to pass. You can retry as many times as you like.

Study the course →

Question 1What is an SLI in SRE practice?
Question 2A service has a 99.9% SLO over 30 days. Roughly how much total downtime does the error budget allow?
Question 3The error budget for the quarter is fully spent and a risky feature release is planned for tomorrow. According to the agreed error budget policy, what should happen?
Question 4How should the SLO relate to the SLA offered to customers?
Question 5Which is the best SLI for a checkout API?
Question 6Which task is a typical example of toil?
Question 7A director demands 99.999% availability for an internal reporting tool that is used a few times a week. What is the best response?
Question 8What best describes the difference between monitoring and observability?
Question 9You need to see which of six microservices adds the most delay to a single slow request. Which signal fits best?
Question 10A developer adds the user ID as a label on a request-count metric. Millions of users exist. What is the main risk?
Question 11Which HTTP header carries the trace ID between services in the W3C standard used by OpenTelemetry?
Question 12Your traces show the first two services but stop at the third, which is a legacy service. What is the most likely cause?
Question 13What is the main benefit of tail sampling in the OpenTelemetry Collector?
Question 14Your logs contain customers’ email addresses and full tokens. What is the best fix that keeps useful debugging data?
Question 15Which alert is best suited to wake someone at night?
Question 16With a 30-day SLO window, a burn rate of 14.4 sustained for one hour consumes about what share of the error budget?
Question 17An alert fired 12 times this month and the on-call engineer took no action each time. What should the team do?
Question 18What is the purpose of a runbook linked from an alert?
Question 19The primary on-call does not acknowledge a page within the set time. What should the system do?
Question 20A team of three people shares one on-call rotation, each taking one week in three plus holidays. Which problem does this most likely cause?
Question 21What is the main job of the incident commander?
Question 22A bad release at 14:00 is causing errors. A rollback is possible in two minutes, but the team wants to understand the bug first. What should the incident commander decide?
Question 23A draft postmortem says “Root cause: engineer made a typo and should be more careful.” What is the best improvement?
Question 24What makes a postmortem action item useful?
Question 25During a long outage, executives keep messaging engineers directly for updates. Which role should handle this?
Question 26Why is “human error” usually not an acceptable final root cause in a postmortem?
Question 27What does “build once, deploy many” mean in a CI/CD pipeline?
Question 28A GitHub Actions workflow must deploy to a cloud account. Which approach avoids storing long-lived cloud keys?
Question 29In a Terraform pipeline, where should the “apply” step run after a pull request is merged?
Question 30Which set lists the DORA throughput metrics?
Question 31A pipeline takes 55 minutes, mostly slow end-to-end tests, so developers batch up large changes. What is the best first improvement?
Question 32A scheduled terraform plan shows changes nobody committed: a security group was edited in the cloud console. What is this called, and what is the right response?
Question 33Why pin a third-party CI action to a full commit SHA instead of a floating tag?
Question 34What is the main advantage of a blue-green deployment?
Question 35You want to expose a risky new payment flow to real traffic while limiting harm if it breaks. Which strategy fits best?
Question 36A new release needs a database column renamed. Which approach keeps rollback possible?
Question 37What do feature flags let a team do?
Question 38A global security rule is about to be pushed to all regions at once with no staged rollout. Based on the Cloudflare 2019 incident, what is the safest change?
Question 39A rollback has never been rehearsed, and a release is scheduled for tonight. What should the team do first?
Question 40Which feature of Argo Rollouts supports automated promotion or abort of a canary?