[GOBBLIN-1914] Configurable Gobblin Application Master class for Yarn

[homatthew]

Dear Gobblin maintainers,

Please accept this PR. I understand that it will not be reviewed until I have checked off all the steps below!

JIRA

• My PR addresses the following Gobblin JIRA issues and references them in the PR title. For example, "[GOBBLIN-1914] My Gobblin PR"
  • https://issues.apache.org/jira/browse/GOBBLIN-1914

Description

• Here are some details about my PR, including screenshots (if applicable):
  The current design of the GobblinApplicationMaster has a heavy dependency on Helix task framework. Going forward, we'd like to explore Gobblin on Temporal as an alternative to Helix TF. I am requesting the ability to add specify a custom Application Master class from the Azkaban yarn app launcher to allow experimentation with different GobblinAM implementations that do not rely on Helix.

Tests

• My PR adds the following unit tests OR does not need testing for this extremely good reason:
  Existing unit tests / internal testing

Commits

• My commits all reference JIRA issues in their subject lines, and I have squashed multiple commits if they address the same issue. In addition, my commits follow the guidelines from "How to write a good git commit message":
  1. Subject is separated from body by a blank line
  2. Subject is limited to 50 characters
  3. Subject does not end with a period
  4. Subject uses the imperative mood ("add", not "adding")
  5. Body wraps at 72 characters
  6. Body explains "what" and "why", not "how"

[codecov-commenter]

Codecov Report

Merging #3781 (2a5ff84) into master (a289613) will increase coverage by 0.00%.
Report is 3 commits behind head on master.
The diff coverage is 83.33%.

@@            Coverage Diff            @@
##             master    #3781   +/-   ##
=========================================
  Coverage     47.30%   47.30%           
- Complexity    10953    10955    +2     
=========================================
  Files          2152     2152           
  Lines         85114    85111    -3     
  Branches       9452     9452           
=========================================
+ Hits          40263    40266    +3     
+ Misses        41197    41193    -4     
+ Partials       3654     3652    -2     

Files Changed	Coverage Δ
...rg/apache/gobblin/yarn/GobblinYarnAppLauncher.java	26.58% <81.81%> (+0.49%)	⬆️
...che/gobblin/yarn/GobblinYarnConfigurationKeys.java	87.50% <100.00%> (+1.78%)	⬆️

... and 10 files with indirect coverage changes

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[phet]

should it really be checked? e.g. how likely is it that a caller would want to handle CNFE specifically?

we might throw something more general, like IOE... or just wrap it in RuntimeException. looking at a few other uses of reflection for dynamic instance-construction, I don't see reflection-based failures showing up in the exception signature

[phet]

these are in the same package, so potentially OK... but just noting this introduces a circular dependency between the two classes

[homatthew]

Worth calling out. I can't think of a scenario where this would be a problem but open to input

[ZihanLi58]

+1