Sobes.tech

Have you been responsible for product operation? Was support handled by your team? Were alerts and metrics routed to you?

198

What happens if one of the consumers in the group goes down? Tell about rebalancing.

198

When does the Garbage Collector run? Are there specific stop points or moments?

197

Tell more about working with Kafka: how exactly message deduplication was performed?

194

What types of replication are there (synchronous and asynchronous)? How does synchronous replication work and what are its drawbacks?

192

Is it possible that messages are unevenly distributed across partitions? How to deal with it?

192

Tell about the team composition, service responsibilities, and release cycle.

190

Why was gRPC chosen instead of REST? Were there specific performance requirements?

190

Abstract case: the service crashes (OOM kill), you see no errors, but the last status in the pod indicates a memory shortage. What is the sequence of actions for diagnostics and stabilization?

189

What is the difference between EXPLAIN and EXPLAIN ANALYZE? Why should EXPLAIN ANALYZE not be used in production?

184

If more consumers are added with a large Kafka lag — will it help? In which cases will it help, and in which will it not? What if there are 16 partitions and 2 consumers?

182

Did you store all messages from Kafka in a separate table? What was the point of this, was Kafka insufficient?

180

The alert triggered when RPS increased to 5000. What did you do in such a case? How did you determine whether it was a real load increase or a system problem?

171

Please tell us what you do at your current job, what you like, which tasks you enjoy, which tasks you dislike, and would like to avoid?

171
/3