Section checkpoint

Key Technologies

10 questions drawn from the 5 chapters in this section. If most land on the first try, you are ready for the next section. If not, the explanation after each answer says which idea to revisit.

Checkpoint

10 questions · first-try answers earn the most XP

1. Redis is configured with appendonly yes and appendfsync everysec. The machine loses power. Roughly how much acknowledged data can be lost?

2. A Redis lock is taken with `SET lock:order:42 <token> NX PX 30000`. The holder pauses for 40 seconds (a long garbage collection), then writes to the database. What can go wrong?

3. A topic has 6 partitions and an order service keys messages by order ID. The team raises it to 12 partitions. What changes for existing orders?

4. A service reads from Kafka, charges a card through a payment API, and writes a result back to Kafka using transactions. Does Kafka's exactly-once make the charge happen exactly once?

5. A team needs messages to survive the loss of one RabbitMQ node in a 3-node cluster. Which queue type fits on RabbitMQ 4.x?

6. Publishers suddenly stall and connections show as blocked. Memory on one node is at 61% of RAM with default settings. What is happening?

7. A standard queue keeps messages 4 days. Its DLQ also keeps them 4 days. A message fails for 3 days, then moves to the DLQ. How long will it stay there?

8. A Lambda function processes SQS batches of 10. One bad message makes the whole batch fail, so 9 good messages are retried again and again. What fixes this?

9. A team embeds Hazelcast members in their web app pods and autoscales the pods every few minutes. What problem is most likely?

10. An IMap uses a MapStore with write-delay-seconds set to 5. What does that mean for the database?

What this checkpoint covers

Put it to work in a drill

Next section: Design Problems