Sdap 268: asynchronous request management

[tloubrieu-jpl]

1. added the fair scheduling option in spark (tested on podaac-devk8s, one big request launched first, small request second, second result comes first, see test code https://github.com/tloubrieu-jpl/sdap_nexus_tests/)

2. added NexusHandlerManager to manage different end-points for a single algorthim. The demonstrate that with:

   • legacy end-points
   • rest end point (for the Spark, for example /TimeAvgMapSpark --> /algorithms/TimeAvgMap )

3. add asynchronous management on the API, after given time (hard code for now), request sends back 202 status and a job id. The job status and eventually the job result can be found from url /jobs/<job_id>

The webapp docker image is available on https://hub.docker.com/repository/docker/tloubrieu/nexus-webapp

4. asynchronous parameters (time out, cleaning period and request validity period) are still hard coded.