Mastering Retry Logic in Java: Spring‑Retry vs Guava‑Retry

Spring‑Retry provides declarative retry support for Spring applications, suitable for batch processing, integration, and Hadoop. After adding the Maven dependency, you can create a task method that randomly throws exceptions or returns values, then use RetryTemplate to configure a SimpleRetryPolicy (max attempts) and a FixedBackOffPolicy (retry interval). The template executes the task via execute, requiring a RetryCallback for the main logic and a RecoveryCallback for fallback when retries are exhausted or a non‑retryable exception occurs.

Retry Policies

NeverRetryPolicy : no retries.

AlwaysRetryPolicy : unlimited retries (risk of infinite loop).

SimpleRetryPolicy : fixed number of attempts (default 3).

TimeoutRetryPolicy : retries within a time limit.

ExceptionClassifierRetryPolicy : different policies per exception type.

CircuitBreakerRetryPolicy : adds open/close timeouts and a delegate.

CompositeRetryPolicy : combines multiple policies (optimistic or pessimistic).

Back‑off Strategies

NeverBackOffPolicy : immediate retry.

FixedBackOffPolicy : fixed sleep period (default 1 s, configurable via sleeper and backOffPeriod).

UniformRandomBackOffPolicy : random sleep between minBackOffPeriod and maxBackOffPeriod.

ExponentialBackOffPolicy : exponential increase using initialInterval, maxInterval, and multiplier.

ExponentialRandomBackOffPolicy : exponential back‑off with random jitter.

When using Spring‑Boot, enable retry with @EnableRetry and annotate methods with @Retryable (specifying exception types, max attempts, and back‑off) and recovery methods with @Recover. Example service code demonstrates a method that retries on RemoteAccessException with a 2‑second initial delay that doubles each attempt.

Guava‑Retry

Guava‑Retry (artifact guava-retrying) offers a thread‑safe retryer built around java.util.concurrent.Callable. Using RetryerBuilder, you can define retry conditions based on exception types ( retryIfExceptionOfType) or result values ( retryIfResult), set a fixed wait strategy, and limit attempts with a stop strategy.

Retryer<Boolean> retryer = RetryerBuilder.<Boolean>newBuilder()
    .retryIfExceptionOfType(RemoteAccessException.class)
    .retryIfResult(res -> res == false)
    .withWaitStrategy(WaitStrategies.fixedWait(3, TimeUnit.SECONDS))
    .withStopStrategy(StopStrategies.stopAfterAttempt(3))
    .build();

The retryer is invoked via retryer.call(() -> RetryDemoTask.retryTask("abc")). Guava‑Retry also supports listeners ( RetryListener) to log each failed attempt.

Comparison

Both frameworks are thread‑safe and suitable for concurrent retry scenarios. Spring‑Retry integrates tightly with the Spring ecosystem and retries based on declared exceptions, while Guava‑Retry provides more flexible criteria, including result‑based retries and richer back‑off configurations. In practice, Guava‑Retry often requires less boilerplate for simple use cases.

Overall, choose Spring‑Retry when you need seamless Spring integration and exception‑driven retries; choose Guava‑Retry when you prefer result‑oriented logic and more granular control over retry policies.