ID: 11247
Title: Preserve piggyback data during restart and updates
Component: Site Management
Level: 1
Class: New feature
Version: 1.7.0i1
For performance reasons, the Piggyback data is stored in the tmpfs of the site
(in tmp/check_mk/piggyback). This has the disadvantage that the files disappear
as soon as the checkmk server is restarted or the tmpfs is unmounted (e.g.
during an update).
The lack of piggyback data can temporarily lead to incorrect check results and,
if you use the Dynamic Configuration, even cause certain hosts/services to
disappear from the configuration for a short time.
With this change the content of tmp/check_mk/piggyback and
tmp/check_mk/piggyback_sources is saved to var/omd/tmpfs-dump.tar during
certain actions (currently: omd stop, omd umount).
If the "omd start" detects that tmpfs needs to be reinitialized, the backup
from var/omd/tmpfs-dump.tar is restored to the fresh tmpfs.
ID: 11126
Title: ucd_cpu_util: Fix wrong CPU utilization
Component: Checks & agents
Level: 1
Class: Bug fix
Version: 1.7.0i1
This fixes a wrong calculation in the check. The CPU utilization used the wrong values.
The rates for each value got calculated twice which resultet in bogus results.
ID: 11085
Title: Icon upload: Add missing transaction validation
Component: WATO
Level: 1
Class: Security fix
Version: 1.7.0i1
The transaction IDs (CSRF tokens) were not validated while processing the upload of icons.
This alone is not a security hole, rather a lack of validation of this call.
ID: 11221
Title: systemtime: Offset for vSphere special agent
Component: Checks & agents
Level: 1
Class: Bug fix
Version: 1.7.0i1
The check <tt>systemtime<\tt> reported an incorrect offset when checking
the output of the vSphere special agent. This was caused by the runtime of
the special agent itself and has now been fixed. The check <tt>systemtime<\tt>
can now process a second, optional input which gives the timestamp against
which the actual offset is computed.
ID: 11046
Title: Fix infinite reloads of views when the number of columns is changed
Component: Multisite
Level: 1
Class: Bug fix
Version: 1.7.0i1
When the reload of a view was set to "off" and the number
of columns was modified this lead to infinite reloads of
the displayed table. This is fixed.
ID: 11246
Title: Drop classic theme
Component: Multisite
Level: 2
Class: New feature
Version: 1.7.0i1
With this change we remove the Classic theme. Even though there are still many
fans of this theme who like to use our Classic Theme, we can no longer support
this theme in the future due to cost and effort.
Users of the Classic Theme will automatically be switched to the current
default theme.
The main reason for this change is that we will make massive changes to the
user interface of Checkmk with the upcoming version. We can easily implement
these changes in the modern and dark theme together. In addition, we would have
had to redesign and implement some of these in the Classic theme, which we
cannot afford.
ID: 11001
Title: Periodic Service Discovery: Add white-/blacklist for removing vanished services automatically
Component: Checks & agents
Level: 1
Class: New feature
Version: 1.7.0i1
Within the ruleset {{Periodic service discovery}} you can indicate service
names or regexes in {{Activate only services matching}} resp.
{{Don't activate services matching}} below
{{Service Filters}} > {{Dedicated white-/blacklists for new and vanished services}}
in order to control which new detected services are added automatically.
Now there are two new lists {{Remove only matching vanished services}} and
{{Don't remove matching vanished services}} for vanished services below
{{Service Filters}} > {{Dedicated white-/blacklists for new and vanished services}}
in order to control which vanished services are removed automatically.
Below {{Service Filters}} > {{Combined white-/blacklists for new and vanished services}}
you can give patterns for both new and vanished services.
Migration notes: Previously there were only the options
{{Activate only services matching}} and {{Don't activate services matching}}
for new {{and}} vanished services. These kind of rules are migrated to be
compatible: these lists can now be found below
{{Service filters}} > {{Combined white-/blacklist for new and vanished services}}.
ID: 11245
Title: Dashboard: Improve context sensitive dashboard handling
Component: Multisite
Level: 2
Class: New feature
Version: 1.7.0i1
Each user opening a dashboard is now able to apply custom filters to a dashboard
when viewing it. This can be done by opening the "Update context" dialog from
the dashboard menu.
>From that menu it is possible to select one or multiple filters that will then
be used to filter the data shown by the dashboard.
This is much like the filter form in the views.
One additional feature has been added that can be used to enforce the user to
use one or multiple filters when initially opening a dashboard.
Imagine you want to have a dashboard that shows data of hosts in a given
hostgroup, for example you want to have a linux, windows, unix dashboard and so
on.
In previous versions you could build one dashboard and clone it multiple times
while changing the filter group name for each of the cloned dashboards. That
would work, but you would end up with a lot of similar dashboards where only
the hard coded host group filter was different.
Now you can define a single dashboard and select the "Host is Group" filter in
the new "Required context filters" option from the dashboard properties. Once
you have done this, the dashboard will automatically open the "Dashboard Context"
dialog when initially opening the dashboard. After selecting the host group, the
dashboard will be rendered.
Btw.: You could also open the dashboard with the prefilled URL parameters to
prevent the "Dashboard Context" dialog from popping up.
ID: 11229
Title: Check_MK discovery: Was not always able to discover new snmp checks
Component: Checks & agents
Level: 1
Class: Bug fix
Version: 1.7.0i1
This werk really fixes the intension of werk #10534:
The {{Check_MK discovery}} check was unable to discover entirely new
check_types. Its discovery phase always relies on cached data, if available.
Since the SNMP datasource only fetches the data it actually needs, there is no
guarantee that all services will be discovered.
So the {{Check_MK discovery}} service failed to discover any interfaces, if the
snmp host did not have any interfaces beforehand.
Through WATO however, the discovery was successfull, since this mechanism may
bypass the snmp caching entirely.