Knowledge Check - Production Operations
Test your understanding of the key concepts from Module 5. Select the best answer for each question, then submit to see your results. You need to score at least 60% to pass.
TaskBoard's Logging Service pipeline has a kubernetes log stream with protocol = "tcp" and the tag k8s. A developer deploys Fluent Bit on MKS, but no logs arrive in Grafana. Which [OUTPUT] section is correct?
A TCP log stream expects Fluent Bit's forward output on port 9000 with TLS, the stream's tag, and the pipeline key as Shared_Key. The Cloud API token authenticates management calls such as creating the pipeline, not log ingestion. logging.<region>.ionos.com/pipelines is the management API, not the pipeline's ingestion address.
During an incident review, a developer runs the Activity Log script from Unit 5.1 with dateStart set 60 days in the past. The response contains entries from the last few weeks only. What explains this, and what should the team do?
The Activity Log keeps entries for 35 days, so a query further back returns nothing older. For a longer audit history, a scheduled job must copy entries to long-term storage such as Object Storage. Paging with limit and offset controls how many entries come back per request, not how far back they reach, and the privilege grants access to the API as a whole.
A scheduled job rotates TaskBoard's CI token. Which order of operations avoids an outage of every pipeline?
Deleting a token deactivates it immediately, and a token's value is shown only once, at generation. So the new token must be verified and stored before the old one is revoked. GET /auth/v1/tokens returns IDs and dates, never token values, and criteria=ALL would revoke every token at once, including the ones other jobs still use.
A developer wants TaskBoard's API pods on Managed Kubernetes to accept traffic only from the ingress controller's namespace. They write this Terraform:
resource "ionoscloud_nsg_firewallrule" "api_from_ingress" {
datacenter_id = ionoscloud_datacenter.taskboard.id
nsg_id = ionoscloud_nsg.k8s_nodes.id
type = "INGRESS"
protocol = "TCP"
port_range_start = 8000
port_range_end = 8000
}
Why does this not achieve the goal, and what does?
NSG members are VMs and NICs, and Managed Kubernetes node pool nodes are excluded, so no NSG rule can isolate pods. A NetworkPolicy with a namespaceSelector for the ingress namespace restricts traffic between pods, and Calico enforces it. NSGs do not bind to the ALB either, because a load balancer is not an NSG member.
TaskBoard is managed by an ArgoCD Application with automated: { prune: true, selfHeal: true }. The application CI still ends with:
kubectl set image deploy/taskboard-api api="$CR_HOST/taskboard/api:$SHA"
After every release, the new image runs for a short time and then the previous image is back. What is the fix?
With selfHeal, ArgoCD treats Git as the desired state and reverts the tag that kubectl set image changed in the cluster. Promoting the tag through a commit to the overlay makes Git and the cluster agree, and the CI job no longer needs cluster credentials. Ignoring the container spec would hide every image change from ArgoCD, and prune only removes objects that were deleted from Git.