All 48 questions

  1. How would you scale workers to keep up with a growing queue backlog?

    Asked in 3 of 217 interviews · last asked Oct 22 · Scenario · Senior

    Show answer

    I would scale workers based on queue depth, message age, and how fast messages arrive and finish. I would add enough capacity to handle new messages and clear the backlog within a target time. I would limit scaling to protect downstream systems, and scale down slowly when demand falls.

    • Track backlog growth and oldest message age, not just worker CPU usage.
    • Use arrival rate and measured worker throughput to estimate the workers needed.
    • In a standard Kafka consumer group, consumers assigned partitions cannot outnumber partitions across subscribed topics.
    • Check downstream limits and processing bottlenecks before adding more workers.

    What the interviewer checks: The interviewer checks whether you can size worker capacity while respecting queue limits and downstream capacity.

    Other ways it was asked
    • “How do you ensure there are enough workers to take things off the queue?”
    • “How would you handle a growing queue backlog when usage increases?”
    • “How should worker instances scale as a queue grows?”
  2. How would you design a distributed task queue for background jobs?

    Asked in 2 of 217 interviews · last asked Sep 1 · Design · Senior

    Show answer

    I would use an API to submit jobs, a durable message broker to hold them, and workers to process them. Workers record results in a database and acknowledge jobs only after success. I would retry failures with delays and move repeatedly failing jobs to a dead-letter queue. Jobs must be safe to repeat because delivery can happen more than once.

    • Run more workers as the backlog grows, but limit concurrency to protect downstream services.
    • Use job IDs to track status and prevent duplicate effects, such as sending the same email twice.
    • Recover unfinished jobs when a worker crashes, using broker-specific acknowledgment or lease rules.
    • Monitor queue depth, job age, failures, and worker health.

    What the interviewer checks: The interviewer checks whether you can separate background work from requests and handle failures, duplicates, and growing load.

    Other ways it was asked
    • “How do you design a distributed task queue for background jobs?”
    • “How would you design a simple event-driven background job for something like sending emails or processing long tasks? Just the high-level components.”
  3. What problems could arise if a notification system sends a batch only when the queue contains more than 10 messages?

    Asked once · last asked Mar 31 · Scenario · Mid-level

    Other ways it was asked
    • “What problem could occur if a notification system waits for a queue to exceed 10 messages before sending, when it receives only two messages per day?”
  4. How would you run dynamic background jobs as Kubernetes Jobs or Rails processes, consume work from a queue, and handle long-running tasks?

    Asked once · last asked Jul 24 · Design · Senior

    Other ways it was asked
    • “How do dynamic background jobs run as Kubernetes Jobs or Rails processes, pick up work from a queue, and handle long-running tasks?”
  5. What data structure ensures fairness in checkout processing, and how do we prevent the payment system from overloading?

    Asked once · last asked Oct 6 · Design · Mid-level

    Other ways it was asked
    • “What data structure would you use to ensure fair first-come-first-served processing for checkout payments while preventing the payment system from being overloaded?”
  6. How do Bitrix agents compare with Laravel queues?

    Asked once · last asked Oct 12 · Concept · Mid-level

    Other ways it was asked
    • “What capabilities do Laravel queues provide that Bitrix agents lack?”
  7. How would you assess a backend service's request capacity and design its queues and concurrency?

    Asked once · last asked Feb 23 · Design · Senior

  8. How would you modify a notification system to batch messages instead of sending them immediately whenever dispatch is called?

    Asked once · last asked Mar 31 · Design · Mid-level

  9. In a batched notification system, how would you decide when to send the queued messages?

    Asked once · last asked Mar 31 · Scenario · Mid-level

  10. During intensive AI training cycles that spike database load and background jobs, how would you design backpressure so APIs stay responsive, using queues, rate limits, circuit breakers, or load shedding?

    Asked once · last asked May 5 · Design · Senior

  11. How would you implement queue prioritization and fairness so interactive requests aren't starved by large AI training jobs, and what metrics would you watch to validate it?

    Asked once · last asked May 5 · Design · Senior

  12. How would you send email to one million users in Laravel reliably, without the process being interrupted?

    Asked once · last asked Jul 16 · Scenario · Senior

  13. How would you process recipient batches and send email data in a Laravel bulk-email workflow?

    Asked once · last asked Jul 16 · Scenario · Senior

  14. How would you track batches in a Laravel bulk-email workflow so that later executions send emails only to recipients who have not yet received them?

    Asked once · last asked Jul 16 · Scenario · Senior

  15. How would you coordinate Laravel commands and queued jobs to process millions of records and send email at scale?

    Asked once · last asked Jul 16 · Design · Senior

  16. How do you choose and configure a queue strategy for a Laravel application?

    Asked once · last asked Jul 16 · Tooling · Senior

  17. If a job keeps failing concurrently, how do you handle the failed jobs and retry them in production?

    Asked once · last asked Jul 16 · Scenario · Senior

  18. How would you configure BullMQ retries and backoff for a 24/7 Zoom recording pipeline encountering API rate limits, and what does fail-safe deletion mean in that architecture?

    Asked once · last asked Aug 2 · Tooling · Senior

  19. What happens to the system and the data if Redis crashes while using Sidekiq?

    Asked once · last asked Aug 26 · Scenario · Senior

  20. How would you handle jobs that hang or fail when a worker machine crashes?

    Asked once · last asked Aug 26 · Scenario · Senior

  21. Should a master job enqueue batches of 100 rows, with each batch then creating another 100 jobs, one per row?

    Asked once · last asked Sep 2 · Design · Senior

  22. In a batched CSV import, should the batch worker insert records before dispatching row-level jobs, or should it delegate insertion and hydration to the final worker layer?

    Asked once · last asked Sep 2 · Design · Senior

  23. When might inserting records in the batch worker before dispatching row-level jobs be a poor choice or fail?

    Asked once · last asked Sep 2 · Scenario · Senior

  24. How would you divide queues or topics for different image types or business scenarios?

    Asked once · last asked Sep 18 · Design · Senior

  25. How would you implement sending an email template to 1,000 leads every Monday?

    Asked once · last asked Sep 28 · Scenario · Mid-level

    Other ways it was asked
    • “How would you handle sending an email and a WhatsApp message to 1,000 leads at the same time every Monday?”
  26. Why would you use multiple named queues instead of a single queue in a background job system?

    Asked once · last asked Sep 28 · Concept · Mid-level

  27. How would you ensure that each lead is processed by only one worker?

    Asked once · last asked Oct 1 · Design · Senior

  28. What should you consider when horizontally scaling a message consumer that processes messages, uses a database, and calls an external API?

    Asked once · last asked Oct 9 · Scenario · Senior

  29. How would you handle backpressure in an asynchronous pipeline with a queue and workers to avoid unbounded memory growth when producers outpace consumers?

    Asked once · last asked Oct 28 · Design · Senior

    Other ways it was asked
    • “How do you handle back pressure in an async pipeline? For example, queue plus workers to avoid unbounded memory growth when producers outpace consumers?”
  30. How do you handle fire-and-forget background work in a Node.js API, for example, sending emails while still ensuring retries, back pressure, and safe shutdown?

    Asked once · last asked Nov 1 · Design · Senior

  31. How would you handle a client that replays a huge backlog after being offline for weeks without overwhelming real-time processing or missing SLAs for fresh events?

    Asked once · last asked Nov 1 · Design · Senior

  32. How would you isolate and throttle replayed offline-event backlogs from fresh events using partitions, queues, consumer groups, or prioritization, and which metrics would drive the controller?

    Asked once · last asked Nov 1 · Design · Senior

  33. How would you design backpressure and end-to-end flow control across queues and downstream workers in an asynchronous pipeline so that overload degrades performance gracefully without compromising correctness?

    Asked once · last asked Nov 10 · Design · Senior

    Other ways it was asked
    • “How would you design back pressure and end-to-end flow control in an async pipeline in queues or workers downstream so overload degrades gracefully without dropping correctness?”
  34. Which Symfony component would you use to handle an asynchronous message queue?

    Asked once · last asked Nov 17 · Concept · Mid-level

  35. Would you run a consumer inside the Go API or deploy it separately, and how would you make that decision?

    Asked once · last asked Nov 17 · Design · Senior

  36. How would you handle a request while its queued work is waiting to be processed?

    Asked once · last asked Nov 19 · Design · Mid-level

  37. What should the user receive in the response before queued work completes?

    Asked once · last asked Nov 19 · Design · Mid-level

  38. What is Celery, and what is it used for?

    Asked once · last asked Feb 23 · Concept · Junior

  39. What is Laravel's queue system?

    Asked once · last asked Mar 12 · Concept · Junior

  40. How would you determine the maximum queue size for a notification batching system?

    Asked once · last asked Mar 31 · Scenario · Mid-level

  41. Where should you run Laravel jobs, and when would you use the CLI, Tinker, or another execution method?

    Asked once · last asked Jul 16 · Tooling · Mid-level

  42. How does Celery work under the hood?

    Asked once · last asked Aug 14 · Concept · Mid-level

  43. Would you perform transaction matching directly in a worker or enqueue a separate job, and why?

    Asked once · last asked Aug 26 · Design · Mid-level

    Other ways it was asked
    • “Does the worker do the transaction matching right away, or does it kick off a new job to handle that?”
  44. How does the asynchronous job pattern work with multiple vendors?

    Asked once · last asked Aug 26 · Design · Senior

  45. What is a queue in a background job system?

    Asked once · last asked Sep 28 · Concept · Junior

  46. What is BullMQ used for?

    Asked once · last asked Oct 7 · Concept · Junior

  47. How do we implement background jobs?

    Asked once · last asked Oct 14 · Design · Mid-level

  48. Which operation in a service would you move behind a queue?

    Asked once · last asked Nov 19 · Design · Mid-level

More Kafka and Message Queues interview questions

How this list is made

Every question here was asked by a real interviewer in a real job interview that a candidate ran MeetAssist in. After the call, the interviewer's questions are pulled from the transcript. Practice sessions, tests, exams and anything that was not a job interview are dropped.

Questions are rewritten so they make sense on their own: names of people, employers and clients are removed, and speech-to-text mistakes are fixed. Small talk, logistics and questions about one candidate's own résumé are left out. Rephrasings of the same question are merged, and the count is the number of different interviews it came up in.

The interviews come from candidates who use MeetAssist, so they lean towards software, data and AI roles: the shares describe these interviews, not every Kafka and Message Queues interview. Answers to the most asked questions are written and checked by an editor; a "common mistake" is shown only when reviews of real answers in these interviews show it, with the count.

217 interviews, 133 candidates, Jan 2026 to Nov 2026. Updated every month with new interviews.