UI: Implement timeouts, retries for jobs

[JasonWeill]

Problem

This should be fixed after #90 (API) is done.

Users should be able to set timeouts and retry options for jobs.

Proposed Solution

Implement these features in the create job UI, mapped to the fields listed below in the API.

• timeout_seconds timeout in seconds for the job to complete (must be a number)
• retry_on_timeout whether to retry if a timeout has occurred (default no)
• max_retries no of times to attempt retry (number)
• min_retry_interval_millis delay in milliseconds between retry attempts (number)

All of these inputs should be optional and should only be sent in the API request if they are present (nonblank/defined).

Additional context

https://github.com/jupyter-server/jupyter-scheduler/pull/45#discussion_r986892626

[aiqc]

Agreed, especially on retries. Sometimes internet dependent processes get hung or a Docker container crashes for unknown reasons.