Figure out a way to make tests more reliable

[yamalight]

Currently the tests for server will randomly fail on Travis-CI with random 500 codes.
Locally they execute normally pretty much 100% of time.
I'm guessing this is related to limited amount of resources during execution in Travis.
We need to figure out a way to make it more reliable.
Different test platform? Less parallelism? Something else?

[yamalight]

Common error in server test:

FAIL test/version.test.js
  ● Should get current and latest versions
    expect(received).toEqual(expected)
    Expected value to equal:
      200
    Received:
      500
      39 | 
      40 |   // check response
    > 41 |   expect(response.statusCode).toEqual(200);
         |                               ^
      42 |   expect(result.server).toBeDefined();
      43 |   expect(result.traefik).toBeDefined();
      44 |   expect(result.latestServer).toBeDefined();
      at Object.toEqual (test/version.test.js:41:31)

[yamalight]

Can this be due to rate-limit from github API side? 🤔

[yamalight]

yeap, that was it. fixed now.