Ensure CI staging run reports errors accurately

[conorsch]

Status

Work in progress

Description of Changes

Fixes #4066

The make ci-go target was always exiting zero, due to the in-line bash
if statement, causing failed CI runs to report as successful. Moved the
logic into a wrapper script. Also trimmed out from of the extraneous
Makefile targets related to CI (and corresponding docs), to remove a bit
of the indirection and make maintenance more straightforward for the
team.

Testing

We'll have to make temporary commits on the WIP PR to confirm that an error in e.g. the provisioning flow results in failed CI. Rough checklist is below. All reports should include URLs to the corresponding CI run satisfying the requested check.

• observe passing CI on this branch with base changes
• append a temporary commit designed to trigger failure in CI (e.g. an Ansible fail task); confirm observation of failure in CI
• revert temporary commit, observe CI passing
• rebase to remove temporary commits

It's possibly we've been missing CI-specific problems such as connection timeouts, so we may need to add retries or 😱 sleeps in order to stabilize.

Deployment

None, CI only.

Checklist

If you made changes to the server application code:

• Linting (make ci-lint) and tests (make -C securedrop test) pass in the development container

If you made changes to securedrop-admin:

• Linting and tests (make -C admin test) pass in the admin development container

If you made changes to the system configuration:

• Configuration tests pass

If you made non-trivial code changes:

• I have written a test plan and validated it for this PR

If you made changes to documentation:

• Doc linting (make docs-lint) passed locally

[kushaldas]

The current changes look okay. @conorsch I am guessing next you will add a failed commit.

[emkll]

• observe passing CI on this branch with base changes: https://circleci.com/workflow-run/269d114d-dda9-47f9-a553-5d94ee4d15ec
• append a temporary commit designed to trigger failure in CI (e.g. an Ansible fail task); confirm observation of failure in CI: https://circleci.com/workflow-run/9b5ecfdb-57e0-48a4-adf4-093a023a7c1e
• rebase to remove temporary commits
• ❗ revert temporary commit, observe CI passing: Did not do that, just appended/dropped a fail commit per the above

[emkll]

Visual diff and functional tests look good to me @conorsch. I've taken the liberty of slightly amending the test plan (not reverting the commit / waiting for yet another CI run), see comment above. Happy to do so in a 2nd review pass if you think it's crucial.
👍 to merge once CI passes again.

[conorsch]

LGTM! Thanks for review and final testing bits, @emkll. Let's watch CI runs in the future—we should see a slight uptick in failures, but those failures should be legitimate. If any flakiness occurs due to e.g. connection timeouts, we'll track that separately.