polly retry policy within given period of time

Question

I have the following code:

var expirationPeriodMilliseconds = 1000;
return Policy
   .HandleResult<EnquiryResult>(message => message.Status == EnquiryStatus.SystemError)
   .RetryForeverAsync() // instead run until expirationPeriodMilliseconds
   .ExecuteAsync(async () => await ratingProvider.Enquire(request, cancellationToken));

Is there any solution for polly to use the retry option for a certain period of time?

Did you check the *other* `.RetryXXX` or `WaitAndRetry` methods ? Retry immediatelly or after a period? If the code retries immediatelly after a server fault the most likely outcome is to get the same fault. — Panagiotis Kanavos, Aug 01 '18 at 14:01
Simply saying if EnquiryStatus.SystemError occurs I want to retry constantly for 1000 milliseconds. — Andrei, Aug 01 '18 at 14:09
You can use multiple policies, eg a Retry and Timeout. This was asked in [this issue from 2016](https://github.com/App-vNext/Polly/issues/184). You can wrap the `RetryForeverAsync` inside a `Timeout` with `myTimeout.Wrap(myRetry)` — Panagiotis Kanavos, Aug 01 '18 at 14:18
More also in this question: https://stackoverflow.com/a/43205503 — mountain traveller, Aug 01 '18 at 16:24

score 0 · Answer 1 · answered Sep 18 '22 at 10:41

Even though it is a later answer it might be helpful for future readers

The policy definitions

var globalTimeoutPolicy = Policy
    .TimeoutAsync<EnquiryResult>(TimeSpan.FromMilliseconds(expirationPeriodMilliseconds));

var retryForeverPolicy = Policy
   .HandleResult<EnquiryResult>(message => message.Status == EnquiryStatus.SystemError)
   .RetryForeverAsync();

var combinedPolicy = Policy.WrapAsync(globalTimeoutPolicy, retryForeverPolicy);

Here we have defined two policies
- One for retrying the same action over and over again if it receives SystemError
- Another to terminate the action after N milliseconds
Finally we have chained/combined the policies into one
- Because the globalTimeoutPolicy has been passed first to the WrapAsync that's why it is the outer policy so this timeout overarches all retry attempts
- If we would pass the globalTimeoutPolicy as the second parameter of the WrapAsync then this timeout constraint would apply for each individual retry attempts

The usage of the chained policies

try
{
    var result = await combinedPolicy.ExecuteAsync(
         async (ct) => await ratingProvider.Enquire(request, ct), cancellationToken);
    Console.WriteLine($"Method has successfully completed with status: '{result.Status}'.");
}
catch(TimeoutRejectedException)
{
    Console.WriteLine($"Method has failed after retrying for {expirationPeriodMilliseconds}ms.");
}

Because your ratingProvider.Enquire method anticipates a CancellationToken that's why we can use Optimistic timeout strategy
- Polly will link together your cancellationToken with the TimeoutPolicy's CancellationToken
If the timeout triggers and all previous retry attempts returned with SystemError then the ExecuteAsync will thrown an TimeoutRejectedException

polly retry policy within given period of time

1 Answers1

The policy definitions

The usage of the chained policies