chore: walk test uses in memory storage

[frrist]

• reduce the requirements for running test

I have noticed this test flakes sometimes. On the off chance this is a race condition within Postgres I have removed it as a dependency of the test opting to use an in-memory storage type instead.

[codecov-commenter]

Codecov Report

Merging #1000 (0ac6275) into master (e917f43) will not change coverage.
The diff coverage is n/a.

@@          Coverage Diff           @@
##           master   #1000   +/-   ##
======================================
  Coverage    35.3%   35.3%           
======================================
  Files          31      31           
  Lines        2090    2090           
======================================
  Hits          739     739           
  Misses       1266    1266           
  Partials       85      85           

[kasteph]

@frrist shouldn't one test for Postgres race conditions anyway? This test might actually be surfacing such an issue (which I think you've also brought up before).

[placer14]

It would be good to understand the race condition a bit more, but I'm in favor of decoupling this from the DB state if we can use memory state instead regardless.