What happens to @Scheduled jobs when a Spring Boot service has multiple replicas?

Designs scheduled work for overlap, multiple replicas, restart, failure, and durable ownership instead of assuming one execution.

1Say itAnswer out loud first

2CompareCompare with Core answer

3RefineRefine with Stronger and Advanced

TL;DR

An @Scheduled method runs in every application replica that enables scheduling. If three instances are running, the same schedule can execute three times. I make the work idempotent and decide explicitly whether parallel executions are acceptable. If the business requires one logical execution, I use a durable scheduler, queue, database claim, or proven distributed lock rather than relying on one application instance.

Full answer below

Practice focus

@Scheduled semantics, scheduler capacity, overlap, replicas, idempotency, locking, durable scheduling, and observability.

Answer first

Answer path

Core answer first, then checkpoints toward a stronger answer.

Start
Core answer
Start here. This is the direct spoken answer to practice first.
Overview
Each running application instance owns its own scheduler, so horizontal scaling can multiply a supposedly singular job.
An @Scheduled method runs in every application replica that enables scheduling. If three instances are running, the same schedule can execute three times. I make the work idempotent and decide explicitly whether parallel executions are acceptable. If the business requires one logical execution, I use a durable scheduler, queue, database claim, or proven distributed lock rather than relying on one application instance.
Checkpoint
Stronger answer
Optional upgrade with examples, trade-offs, and sharper phrasing.
Fixed rate can try to maintain a cadence, while fixed delay waits after completion; neither guarantees exactly once. A long run, scheduler pool size, or multiple declarations can create overlap. I persist job identity and progress when work must survive restart, use a clock and time zone deliberately, and record start, finish, outcome, duration, and skipped or duplicate claims. The job's downstream calls still need timeouts and retry-safe behavior.

Review signals

Use these to catch weak phrasing, shallow reasoning, and risky claims.

Common mistakesCandidateCommon mistakesMistakes that weaken the answer or signal shallow understanding.Open details

Details

Assuming @Scheduled runs once for the whole cluster.
Calling it exactly-once execution.
Keeping no record of missed or failed runs.

What happens to @Scheduled jobs when a Spring Boot service has multiple replicas?

Practice focus

Answer path

Overview

Review signals

Follow-up prompts

Continue the section

Tech stacks

Skill areas

Suggest an improvement

Found this answer useful?