Icinga Classic UI 1.13.4 (as offered by Debian 9, replacing Nagios 3 from Debian 8) is exhibiting a strange behaviour when reviewing alert history.
When selecting "View Alert History For This Host" or "View Alert History For This Service" you receive the log entries for today.
After clicking "Latest Archive", you should go back one day, but instead, you are given all log entries from 1970 to yesterday at once.
Under "Log Navigation", the start time is shown as the UNIX epoch, which from looking at the source code, indicates that ts_start is getting zeroed somewhere, or is never being received by the CGI at all. ts_end works correctly, so this implies that ts_start and ts_end are both read from the CGI parameters successfully, so something seems to be zeroing ts_start.
It seems to be too blatant of a fault to be a bug in Icinga, and there seems to be no indication that anyone else has ever experienced this problem. The source code suggests that the time processing is self-contained and is not dependent on configuration or anything, so that would seem to rule out it being a problem moving from Nagios to Icinga.
Any ideas?
(Ah, looking at a stale Nagios tab, I can see that Nagios used archive=n instead of ts_start/ts_end, so this problem arises from the new method of time ranges.)
