Incorrect "Off-Nominal planning: down USS" time ranges

[BenjaminPelletier]

Observed behavior
The "Off-Nominal planning: down USS" appears to have two bugs:

1. (most critically) uss_qualifier-managed op intent is not removed under some conditions
2. The "Verify area is clear" test step does not seem to check the correct time range

For 1, see s328.html, run with v0.25.0. In e15, uss_qualifier creates a reference to op intent 6659 in the DSS. In e27, we can see it's still there. In the cleanup, e34 seems like it should find the reference we need to clear, but instead it returns nothing because the search time range was 16:52:20 to 18:22:21 while 6659's time range was 18:46:47 to 19:31:48. This likely relates to 2 because e12 shows the same time range on the "Verify area is clear search", but this time range is not the relevant time range for the scenario -- the scenario used a much later time range.

I have not attached s329, but it and all future instances of "Off-Nominal planning: down USS" fail because 6659 is still present, "Verify area is clear" passes because the time range queried doesn't match what will be used in the test, and the test doesn't work because the area isn't clear (6659 is present).

Test check

• Successful operational intents cleanup
• Area is clear of op intents

Difference from expected behavior
A test scenario should absolutely not leave the test environment dirty when run to completion, regardless of paths taken within the scenario. So, 6659 should have been deleted within the test scenario.

A test scenario requiring a clear area should check the area it's actually going to use, not an area from an earlier time. I suspect this problem may be due to attempting to evaluate the 4D usage area at the time of test scenario instantiation rather than the time of test scenario execution. This is not a good practice because the 4D usage area is likely to be substantially different at execution time than at instantiation time if the usage area involves "now" in any way (which is very common).

Additional information
I believe "Off-Nominal planning: down USS with equal priority conflicts not permitted" is likely to have a similar problem and should be fixed preemptively if so.

[mickmis]

After some investigations, I believe the most probable root cause for this issue is that we determine the time bounds of the 4D extents in the __init__ constructor of the scenarios rather than in the run function. And I am guessing that the test run that you shared an extract of has a pretty long runtime, at least 1h50? (18:46:47 - 16:52:20). This is consistent with the fact that the run has at least 328 scenarios being ran :)
This is making the 'intents extents' we determine from validate_flight_intent_templates invalid by the time the scenario is actually ran. This function does account for a maximum of 45mins test run duration.
This probably doesn't surface in our CI since we do not have such lengthy runs by far.

If this hypothesis is correct, a relatively simple fix would be to move the determination of the intents extents within the run method. I will give it a shot.