Fix: #254 by increasing timeout, use backoff module instead of flaky

[nicain]

Here I have moved away from "flaky" in favor of "backoff", which was already a dependency of the library.

This repository currently uses the "flaky" module/plugin to pytest to allow retries on tests. As I was trying to quantify this flakiness, I quickly realized that under-the-hood, this module uses runtime modification (monkey-patching) of the usual pytest callbacks, which breaks the interface contracts that the pytest ecosystem subscribes to and does not let me aggregate the timing statistics I need for quantifying flakiness.

The "backoff" module is already in this repo. In this PR I modify the "usual" way we set up fixtures to use the callback system from "backoff".

Fixes #254 by increasing timeout via backoff with 4 tries: [100, 250, 300, 500]

The worst execution time was 234 seconds, so the max allowed is now 500:
Flakiness report:

SUCCESS
       when    duration outcome  trial
0     setup    1.524859  passed      0
1      call  141.551338  passed      0
2  teardown    0.000545  passed      0
       when    duration outcome  trial
0     setup    1.418313  passed      1
1      call  234.604837  passed      1
2  teardown    0.000532  passed      1
       when    duration outcome  trial
0     setup    1.519378  passed      2
1      call  124.333620  passed      2
2  teardown    0.000748  passed      2
       when    duration outcome  trial
0     setup    1.569923  passed      3
1      call  129.542894  passed      3
2  teardown    0.000723  passed      3
       when    duration outcome  trial
0     setup    1.459673  passed      4
1      call  125.115383  passed      4
2  teardown    0.000780  passed      4
       when   duration outcome  trial
0     setup   1.515023  passed      5
1      call  79.950361  passed      5
2  teardown   0.000626  passed      5
       when   duration outcome  trial
0     setup   1.526374  passed      6
1      call  82.278016  passed      6
2  teardown   0.001218  passed      6
       when    duration outcome  trial
0     setup    1.533906  passed      7
1      call  107.097216  passed      7
2  teardown    0.000794  passed      7
       when    duration outcome  trial
0     setup    1.417850  passed      8
1      call  136.093182  passed      8
2  teardown    0.000509  passed      8
       when    duration outcome  trial
0     setup    1.433857  passed      9
1      call  134.896833  passed      9
2  teardown    0.000841  passed      9

[nicain]

@aribray Ping, PTAL?

[aribray]

Looks good to me!

[gcf-merge-on-green]

Merge-on-green attempted to merge your PR for 6 hours, but it was not mergeable because either one of your required status checks failed, one of your required reviews was not approved, or there is a do not merge label. Learn more about your required status checks here: https://help.github.com/en/github/administering-a-repository/enabling-required-status-checks. You can remove and reapply the label to re-run the bot.