Do not anchor an MC to a part of the run that had no collisions

[sawenzel]

This fixes the interaction rate of anchored jobs sampled near end of run.

• Collisions stop before data taking does. In run 571043 the FT0 vertex input reads zero over the last 256 s.
• That zero was forwarded as "-interactionRate 0", so workflow creation aborted and the job ended as ERROR_V.
• The end of the run is now the last scaler interval that counted, mirroring the ITS ramp-up adjustment of the start.
• A timestamp with no collisions is marked excluded, which anchorMC.sh handles. No rate is invented for it.
• AggregatedRunInfo cannot give this; its end of run is the end of data taking.
• The raw rate now comes from o2::ctp::CTPRateFetcher rather than a fixed scaler index, which is not the same class in every run.
• The pile-up correction is guarded against zero colliding bunches and against a rate with no Poisson inverse.

https://its.cern.ch/jira/browse/O2-7143

[github-actions]

REQUEST FOR PRODUCTION RELEASES:
To request your PR to be included in production software, please add the corresponding labels called "async-" to your PR. Add the labels directly (if you have the permissions) or add a comment of the form (note that labels are separated by a ",")

+async-label <label1>, <label2>, !<label3> ...

This will add <label1> and <label2> and removes <label3>.

The following labels are available
async-2023-pbpb-apass4
async-2023-pp-apass4
async-2024-pp-apass1
async-2022-pp-apass7
async-2024-pp-cpass0
async-2024-PbPb-apass1
async-2024-ppRef-apass1
async-2024-PbPb-apass2
async-2023-PbPb-apass5