// use a ConstantBackoff
const constDelayRetry = Policy.handleAll()
.retry()
.attempts(3)
.backoff(new ConstantBackoff(10, 3));
constDelayRetry.onRetry(({ delay }) => console.log('delay for next call: ', delay));
await constDelayRetry.execute(() => {
throw new Error('exception');
});
// use a DelegateBackoff
const delegatedDelayPolicy = Policy.handleAll()
.retry()
.attempts(3)
.backoff(new DelegateBackoff(() => 10));
delegatedDelayPolicy.onRetry(({ delay }) => console.log('delay for next call: ', delay));
await delegatedDelayPolicy.execute(() => {
throw new Error('exception');
});
The output looks like
delay for next call: 10
delay for next call: 10
delay for next call: 10
delay for next call: 0
delay for next call: 10
delay for next call: 10
Conceptually, I would expect all delays are 10, but unfortunately, the first delay of DelegateBackoff is always 0.
Looks like the ExponentialBackoff has the same problem.
But Polly doesn't has such problem.
It looks like RetryPolicy.execute should adjust the logic to backoff.next() first then do the delay.
Let's take look at an example
The output looks like
Conceptually, I would expect all delays are 10, but unfortunately, the first delay of
DelegateBackoff
is always 0. Looks like theExponentialBackoff
has the same problem.But Polly doesn't has such problem. It looks like
RetryPolicy.execute
should adjust the logic tobackoff.next()
first then do the delay.