FT: Retry policy not reporting original Exception (cause) on timeout

Environment Details

• Helidon Version: 2.3.0
• Helidon MP
• JDK version: 11
• OS: Linux
• Docker version (if applicable):

---

Problem Description

When an exception occurs with Retry FT policy and the processing time surpasses the overall timeout, the original cause is never reported.

Steps to reproduce

       Retry.builder()
                .retryPolicy(Retry.JitterRetryPolicy.builder().build())
                .overallTimeout(Duration.ofSeconds(1))
                .build().invoke(() -> CompletableFuture.runAsync(() -> {
                    try {
                        Thread.sleep(2000);
                        throw new RuntimeException("Hello");
                    } catch (InterruptedException e) {
                        e.printStackTrace();
                    }
                })).exceptionallyAccept(Throwable::printStackTrace).toCompletableFuture().join();

[spericas]

Check out #3449.

[martin-sladecek]

The timeout will be triggered all the time not if an exception is thrown.

If this should be the case than this is not how the policy currently behaves. Try removing the runtime exception from above and raising the sleep to like 5s - the execution will go on and finish successfully even though the execution time surpasses the overall timeout by a lot.

However, my point was that in case the execution does end up with an exception, but it happens after the overall timeout is reached, the original exception is gone and will never see it in the logs. I want to see the original exception no matter what happens.