if i have one host (with a service ping), i also have 15min difference (I use same check command for both host status and PING service status)
my delay between SOFT and HARD is 15min btw

if you need any more infos, feel free (y)

just to let you know wolfgang put a solution in https://monitoring-portal.org/index.php?thread/35228-host-in-down-state-is-showing-wrong-duration/

i will give it a try with next upgrade and let you know.

Regards,
Zibfr

Hi there,

the following patch (from @wolfgang) works perfectly:

in base/check.c
Change

	/* update last state change times */
	hst->last_state_change = current_time;
	if (hst->state_type == HARD_STATE)
		hst->last_hard_state_change = current_time;

to

	/* update last state change times */
	if (hst->last_state != hst->current_state)
		hst->last_state_change = hst->last_check;
	if (hst->state_type == HARD_STATE)
		hst->last_hard_state_change = hst->last_check;

I tested this workaround on my test server.

In order to correct this on both our production icinga servers (as my boss only accepts official source code with no correction of any kind), could you please correct this bug ideally for v1.9.2 and v1.12.2 (and/or at least latest 1.13.x version) plz ?

Regards
Zibfr

I'd merge a PR of yours, and you can backport the patch into the old EOL releases then. Icinga 1.x doesn't receive updates anymore, just security related and minor things for current 1.14.