Retry policy guidance

[matthewkmayer]

💡 Feature description

https://docs.rs/dynomite/0.10.0/dynomite/retry/enum.Policy.html describes the default retry policy:

A Default impl of retrying 5 times with an exponential backoff of 100 milliseconds

While the default works great, it'd be awesome to understand why those settings were chosen. Since I don't have experience using DynamoDB with heavy load, I don't know if five retries is a lot or not.

Also I'm rusty (no pun intended) on exponential backoff but the default seems massive for how long a request could wait to retry. While I poked around the source of again I am still not super clear on how it behaves. Perhaps some examples in either that repo or this one of what the backoff does, similar to this but with correct delay numbers 😉 :

1. Retryable error happens
2. Wait 100ms and resend request
3. Retryable error happens again
4. Wait 1,000 ms and resend request
5. Retryable error happens again
6. Wait 1,000,000 ms and resend request