Certified Argo Project Associate (CAPA) free question 10 / 10 · Domain 4: Argo Events
A Calendar EventSource is configured with spec.replicas set to 3. What does Argo Events do with these replicas?
- One Pod serves traffic and the other two stand by, with a standby elected active if the old one is gone
- Three Pods serve traffic at the same time, and the Pods produce duplicate copies of the events
- The first Pod runs, and the other replicas stay unscheduled until the replica number is lowered
- The Pods split the schedule between them, and the events of the calendar are shared out in turn
Show the answer and explanation
Answer: A
Why this is correct
The EventSource controller creates a Kubernetes deployment for each EventSource, with one replica by default, and high availability is achieved with spec.replicas greater than 1. The strategy depends on the type. Calendar is listed under Active-Passive, so only one Pod serves traffic and the rest stand by. One of the standby Pods is automatically elected to be active if the old one is gone. The page also tells you not to scale the replicas up manually.
Why the other options are wrong
"Three Pods serve traffic at the same time, and the Pods produce duplicate copies of the events" … That is the Active-Active strategy, which applies to types such as Webhook, GitHub and AWS SNS. Calendar uses Active-Passive.
"The Pods split the schedule between them, and the events of the calendar are shared out in turn" … The page does not describe a split of the work. With Active-Passive only one Pod serves traffic.
"The first Pod runs, and the other replicas stay unscheduled until the replica number is lowered" … The standby Pods run. One of them is automatically elected to be active if the active Pod is gone.
Key points
Active-Active (Webhook, GitHub, SNS and others) vs Active-Passive (Calendar, Kafka, Minio and others).