K8S API error resilience

## New feature Could Nextflow be a little bit more resilient to transient K8S API issues, e.g. ``` Caused by: Request POST /apis/batch/v1/namespaces/ns-<NAMESPACE>/jobs returned an error code=500 { "kind": "Status", "apiVersion": "v1", "metadata": { }, "status": "Failure", "message": "Internal error occurred: resource quota evaluation timed out", "reason": "InternalError", "details": { "causes": [ { "message": "resource quota evaluation timed out" } ] }, "code": 500 } -- Check '.nextflow.log' file for details ``` Perhaps some sort of configurable retry number for 4xx/5xx error codes? The k8s API can be a little flakey under high load situations, especially when there are a lot of jobs sitting around spamming the API with requests to make pods!