NetflixOSS » SimianArmy » SimianArmy-pull-requests #8 SUCCESS
This pull request looks good

SimianArmy-pull-requests #180 SUCCESS
This pull request looks good

Did you consider using the Amazon ClientConfiguration class instead of rolling custom retry code?

I did something similar on a fork to handle API limits with SimpleDB. The code is simpler and it is globally applied instead of having to roll custom retry code at every client interaction. In my case I just increased the retry count.

ClientConfiguration config = new ClientConfiguration();
configuration.setMaxErrorRetry(11);
....
client = new AmazonSimpleDBClient(configuration);

I experimented with custom retry policies but increasing the retry count worked just as well and was less code.

No; I did a quick and dirty hack for a limit I was seeing in production use :)

However it's a good idea. I'll look into it; and have a look at your latest commit. ebukoski@97c1a3a

AWS explain that you should implement exponential back-off when you receive the server (5xx) or throttling errors.
http://docs.aws.amazon.com/general/latest/gr/api-retries.html
I'll close this and re-implement globally.