ID: 7357
Title: wmi.include: Check_MK Discovery does not report 'vanished services' for WMI services if WMI timeout
Component: Checks & agents
Level: 1
Class: Bug fix
Version: 1.5.0p14
If a timeout occurs in a Windows agent section while executing WMI queries the
related services were handled as vanished services by the Check_MK Discovery
service. This may be misleading because this timeout is short-lived.
The WMI query timeouts are a known problem in Windows and can not be solved by
the Check_MK agent. Please see
https://github.com/martinlindhe/wmi_exporter/issues/89 for further information.
Now the Check_MK Discovery does not take WMI timeouts into account. Thus the
Check_MK Discovery does not report "vanished services" any more. The Check_MK
service handles services with WMI timeouts as "stale" further on.
ID: 7387
Title: apache_status: Ignore certificates from local hosts
Component: Checks & agents
Level: 1
Class: Bug fix
Version: 1.6.0i1
Previously not all cases of failing certificate verfication for
local hosts were handled correctly, which could lead to missing
data for that host.
Instead of handling all cases of invalid certificates we now
omit certificate verification for local addresses entirely.
This only affects the addresses "127.0.0.1", "[::1]", and
"localhost".
ID: 7239
Title: Service discovery: Fix creating disabled services rule for services with special characters
Component: WATO
Level: 1
Class: Bug fix
Version: 1.6.0i1
Services with e.g. "[" in their names could not be disabled using the service discovery page.
ID: 7130
Title: Fixed "Clustered services for overlapping clusters" rule while checking
Component: Checks & agents
Level: 1
Class: Bug fix
Version: 1.6.0i1
Normally each node is part of a single cluster host in Check_MK. Using the
ruleset {{Clustered services for overlapping clusters}} you can assign services
found on a node to different independent cluster hosts, so called overlapping
clusters. This was unregarded while executing clustered services checks on the
related clusters. This may result in wrong check results.
Example:
Each of the two clusters {{cluster1}} and {{cluster2}} has the nodes {{node1}}
and {{node2}}. On each node we have discovered a process {{Process ABC}}.
Within the ruleset {{Clustered services for overlapping clusters}} we assign
the service {{Process ABC}} from node {{node1}} to {{cluster1}}.
Now the service {{Process ABC}} which is assigned ot the cluster {{cluster1}}
processes the data of both nodes and may return a wrong check result in the
moment where {{node1}} AND {{node2}} have data for {{Process ABC}}.
ID: 7386
Title: docker_node_disk_usage: Handle busy docker daemon
Component: Checks & agents
Level: 1
Class: Bug fix
Version: 1.6.0i1
If the docker daemon cannot get the disk usage info, the corresponding services
will become stale without the Check_MK Discovery service going into a WARNING
state.
This can happen if you are calling the mk_docker plugin from different Sites.
Consider configuring asynchronous plugin execution in this case.
ID: 7385
Title: check_http: Fix port mixup in URL-mode
Component: Checks & agents
Level: 1
Class: Bug fix
Version: 1.6.0i1
The check_http check in URL mode mixed up the servers port and the proxy port if a proxy was used.
If users were using a proxy to check a URL, the configured port was used as the proxys port, while
it was not possible to configure a port for the server. Users who have been relying on this wrong
behaviour need to edit their rules.
The rule has been reworked to clearly show which settings apply the server, and which to the proxy.
ID: 7414
Title: mail: fix bug with multiple recipients and synchronous delivery
Component: Notifications
Level: 1
Class: Bug fix
Version: 1.6.0i1
If Email notifictations were sent to multiple recipients and synchronous
delivery via SMTP was used mail servers may process the recipients of
Emails wrong. This may result in failed notifications or only partially
delivered Emails. The exact behaviour depends on the mail server.