status totals are wrong when searching
|Target version:||Icinga 1.x - 1.7|
|Icinga Version:||1.10.0||OS Version:||any|
When a search is done, the totals in status.cgi are all wrong, all numbers are 0, regardless of what is searched for.
classic ui: Added multiple hosts/services to status.cgi GET #1981
This is pretty much a rewrite of status.cgi.some changes:
- you can define more then one object to display
- same with host groups and service groups
- hostgroup=all now only shows which are in host groups
- host sorting is working now
- added 2 new cgi vars
- sortobject = determine if to sort host or service list (host_service_view)
- allunhandledproblems = to display ALL unhanded problems
- fixed and reorganized the links in status totals
- searching returns more then one host/service group if group name matches
- combined Host/Service view now properly supported
- code cleanup
- more little fixes and enhancements
classic-ui: added service / host properties to links in status totals
- now serviceprops and hostprops are added to links in status totals if they were set in the request
- let's see how this is doing
classi-ui: fixed performance issues and other problems in status.cgi
- displaying of host and service group should be much faster now
compared to the speed aftere the "big rewrite"
- fixed filtering for servicegroups
- fixed problem with wrong error message when filters are set
#2 Updated by dnsmichi over 3 years ago
- Target version changed from 1.5 to 1.6
this is introduced when using host filters (precompiled regexes) and the asterisk for searching. search is returning only the first result by default, where the counting of the host and service totals is correct.
in order to fight the error, one might rewrite the search completely, and even if using regex, doing the counts for the status totals right in place (global variable instead of extra counting!).
this relates to #844
#4 Updated by ricardo over 3 years ago
- Estimated time deleted (
Just wanted to know what these counters should represent. because sometimes they counted ALL hosts and services and sometimes just the ones which are displayed.
As we now have the TAC header I think it would make sense just to display the counts for the hosts and services which are displayed.
What do you think?
#5 Updated by dnsmichi over 3 years ago
basically, those counts do the following
- add urls to different states
- host states with servicestatustype=29 (up=2, down=4, unreachable=8, pending=1)
- service states with hoststatustype=15 (ok=2, warning=4, unknown=8, critical=16, pending=1)
- all problems
- hosts states with servicestatustype=29 (down+unreachable+pending = 4+8 = 12)
- service states with hoststatustype=15 (warning+unknown+critical = 4+8+16 = 28)
- all types
- hosts states with servicestatustype=29
- service states with hoststatustype=15
the counts mostly did not happen correctly due to the navbarsearch filter not being honored when counting. having rewritten that, i think they would remain "self corrected" now by fixing #844
#7 Updated by ricardo over 3 years ago
- Subject changed from When searching, totals in page is wrong to status totals are wrong when searching
- Assignee set to ricardo
- Target version set to 1.7
- % Done changed from 0 to 90
in current rbartels/cgi
can't guarantee that all display_styles are displayed correctly, so please TEST!!!
#9 Updated by ricardo almost 3 years ago
- File statustotal_all_hostgroups.png added
Hope people understand the way it works now.
For example: If you display all hostgroups, it only shows the status for all hostgroups, not for host's which aren't in any hostgroup. The same with servicegroups.
example for all hostgroups (4 hosts), filtered for services in unknown state