Cheat sheet & self-check

Pick an answer to see if you got it, and why.

1. Q1. Why relax the chart's ServiceMonitor selector in a lab?

   1. AFor speed
   2. BBy default the chart's Prometheus only selects monitors/rules with its release label; setting the NilUsesHelmValues options to false makes it pick up podinfo's ServiceMonitor and your rules without extra labels
   3. CGrafana needs it
   4. DIt disables alerts
   Show answer

   B. The classic 'my ServiceMonitor is ignored' problem (see Scaling Prometheus, lesson 01).

   From lesson 05 · Observability stack with SLO alerts

2. Q2. What does a fast-burn SLO alert detect?

   1. AHigh CPU
   2. BErrors consuming the error budget much faster than sustainable (e.g. 14.4× over 1 hour and 5 minutes for a 30-day SLO)
   3. CSlow disks
   4. DPod restarts
   Show answer

   B. It pages on user impact at a rate that matters, not on raw thresholds.

   From lesson 05 · Observability stack with SLO alerts

3. Q3. How do you prove the alert pipeline works end to end?

   1. ARead the YAML
   2. BGenerate errors until the rule fires, see it in Prometheus and Alertmanager, and confirm the receiver actually gets the notification
   3. CRestart Prometheus
   4. DDelete podinfo
   Show answer

   B. Test the whole path: metric → rule → Alertmanager → receiver.

   From lesson 05 · Observability stack with SLO alerts

4. Q4. Why can't Argo CD in the mgmt kind cluster use the default kubeconfig address for edge-01?

   1. AArgo CD can't do multi-cluster
   2. Bkind kubeconfigs point at 127.0.0.1:<port> on the host, which from inside the mgmt cluster is the mgmt node itself; use the edge control-plane container's IP on the kind Docker network
   3. CCertificates are missing
   4. DThe port is blocked by Argo CD
   Show answer

   B. Inside a container, localhost means the container, not the host.

   From lesson 06 · A GitOps fleet on kind

5. Q5. You add edge-04 with the label env=edge. What happens?

   1. ANothing until you write new YAML
   2. BThe ApplicationSet's cluster generator matches it and creates the add-on and app Applications for it automatically
   3. CAll clusters resync from scratch
   4. DArgo CD rejects it
   Show answer

   B. Labels become the interface to the fleet.

   From lesson 06 · A GitOps fleet on kind

6. Q6. How do you roll a change to the canary cluster first?

   1. AApply it by hand there
   2. BPoint canary-labelled clusters at the new revision (e.g. main) and stable clusters at the previous tag; promote by moving the stable pointer after the canary is healthy
   3. CDelete the stable clusters
   4. DUse a bigger cluster
   Show answer

   B. Waves are just different revisions per label group, promoted through Git.

   From lesson 06 · A GitOps fleet on kind