Investigate APC data not getting removed from backlog

[mjaakko]

It seems that APC data can get stuck on Pulsar backlogs. Maybe the sink is not acknowledging all messages as it should?

Example log message (see received_at):

{"timestamp":"2023-04-27T13:34:31.200Z","level":"WARN","thread":"main","logger":"fi.hsl.transitlog.MessageHandler","message":"Timestamp (tst) of APC message from vehicle 22/1229 was outside of accepted range. Tst: 1767222445000 (2025-12-31T23:07:25Z), received at: 1675841845992 (2023-02-08T07:37:25.992Z)","context":"default"}