Retry batches on failure

[vishalbollu]

Description

Support retrying of failed batches. Allow users to configure the number of allowed retries for batches in the Job submission request.

POST <batch_api_endpoint>:
{
    "workers": <int>,         # the number of workers to allocate for this job (required)
    "retries": <int>,         # the number of retries for a batch before considering it be failed
    "item_list": {
        ...
    }
}