make db inquiry sequencially

[rhaegar325]

Summary

Order the legacy monitor_requests table by variable name instead of by
requested_at, so the monitor's merged request queue no longer depends on which
second each request landed in.

Changes

TaskTracker.take_monitor_requests — tracking.py

- "ORDER BY requested_at, variable",
+ "ORDER BY variable",

requested_at defaults to CURRENT_TIMESTAMP, which resolves to whole seconds.
Requests enqueued in the same second therefore tied and fell back to the variable
name, and only requests that happened to straddle a second boundary came back in
arrival order — so the ordering could not deliver the arrival order it looks like
it delivers, and test_monitor_requests_are_taken_once failed under load when the
enqueues straddled a tick.

Variable-name order is also what the file queue this table is the legacy half of
already does: task_status.take_monitor_requests reads its requests in filename
order. The monitor concatenates the two, so ordering them the same way is what
makes the merged queue's submission order predictable. The requested_at column
is kept for diagnostics.

New regression test — tests/unit/test_tracking.py

test_monitor_requests_order_does_not_depend_on_the_clock writes the two
timestamps a second apart directly rather than waiting for a boundary, so the
case that used to flake is now covered deterministically.

Impact

Only the order in which variables appended mid-run (moppy-cmorise --append-variables) are submitted as PBS jobs, and only for requests that
straddle a second boundary. Which variables run, and what they produce, are
unchanged. No schema change.

Verification

• The new test fails without the ORDER BY change and passes with it.
• The real flake reproducer (two enqueues either side of a sleep(1.1)) is now
  stable.
• 2085 unit tests pass.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 78.6%. Comparing base (c10597b) to head (ef27188).

Additional details and impacted files

@@          Coverage Diff          @@
##            main    #683   +/-   ##
=====================================
  Coverage   78.6%   78.6%           
=====================================
  Files         41      41           
  Lines       8998    8998           
  Branches    1682    1682           
=====================================
  Hits        7073    7073           
  Misses      1591    1591           
  Partials     334     334           

Flag	Coverage Δ
unit	78.6% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.