ID: 15205
Title: Host labels for AWS EC2 and Azure VM hosts, account label for all AWS hosts
Component: Checks & agents
Level: 1
Class: New feature
Version: 2.2.0i1
Hosts representing Azure Virtual Machines/AWS EC2 instances will now be labelled as such, specifically:
LI: AWS EC2s: "cmk/aws/ec2:instance"
LI: Azure Virtual Machines: "cmk/azure/vm:instance"
Additionally, all AWS hosts will receive a label with the account they belong to: "cmk/aws/account:<account>".
ID: 15261
Title: Changed host label for Azure Resource Group hosts
Component: Checks & agents
Level: 1
Class: New feature
Version: 2.2.0i1
This werk affects users who use the "resource_group:<group>" host label of Azure Resource Group hosts.
The label is being renamed to "cmk/azure/resource_group:<group>" in order to stay consistent with other host labels. The old label will be removed in a future version.
ID: 15234
Title: Periodic service discovery: Vanished clustered services can now be removed automatically
Component: Checks & agents
Level: 1
Class: Bug fix
Version: 2.2.0i1
Up to this Werk vanished clustered services could never be removed using the periodic service discovery.
We have now added a configuration option that allows users to achieve just that, but beware:
By default we keep a record of vanished services on the node if they are assigned to a cluster.
When a clustered service switches from one node to another, it might not be seen on either node for one check cycle.
Keeping clustered services indefinitely keeps us from loosing them in this case.
However this means that truly vanished clustered servces will never be removed from the cluster.
If you choose to include clustered service in the removal operation, vanished services will be removed from clusters, at the risk of loosing services due to the described race condition.
If you have specific needs, you can always adapt the services according to your needs manually using the service discovery page.
ID: 14180
Title: Improved performance when processing clusters and cluster rules
Component: Setup
Level: 1
Class: New feature
Version: 2.2.0i1
Even in configurations with a small number of clusters, you may have experienced performance issues with
<tt>Activate changes</tt> and general checking because cluster service rules were not processed properly.
ID: 15235
Title: Missing agent sections in rare upgrade scenario
Component: Checks & agents
Level: 1
Class: Bug fix
Version: 2.2.0i1
This fixes missing agent sections when users upgrade from a <b>1.6.0 systemd</b> Checkmk agent package to a <b>2.0.0 xinetd</b> package.
Since from 1.6.0 to 2.0.0 we changed the default network service <b>from xinetd to systemd</b>, this is a very rare scenario.
In case it did happen the systemd socket was not stopped during the upgrade, preventing the xinetd service from binding to the port.
This resulted in a partially working monitoring (as the systemd socked <i>was</i> running).
However, services using chached data would go to stale (e.g. <i>"NTP Time"</i>) and the agent updater would no longer be triggered.
ID: 14970
Title: mk_postgres.py: Use PATH as fallback for psql binary location
Component: Checks & agents
Level: 1
Class: Bug fix
Version: 2.2.0i1
This werk affects the monitoring of one or more PostgreSQL instances via agent plugin on linux if the changes made in werk #14975 are applied.
The agent plugin will look for the "psql" binary under the path <tt>/<pg_database>/<pg_version>/bin/<binary_name></tt>, using the configured values in the .env and .pgpass files.
However, this path does not have to be correct, depending on the setup. In case it cannot be found the binary location from PATH is now used instead.
In order to apply this change you will need to reinstall the agent plugin.
ID: 15262
Title: mk_postgres.py: Improved instance process detection
Component: Checks & agents
Level: 1
Class: Bug fix
Version: 2.2.0i1
This werk affects the monitoring of one or more PostgreSQL instances via the Postgres agent plugin for Linux. The changes to the process filtering introduced in werk #15085 could lead to an empty "postgres_instances" section if the instance name contained capital letters.
Now, a case-sensitive filter is applied first, allowing case-insensitive filtering only if no processes could be found.
In order to apply this change you will need to reinstall the agent plugin.
ID: 15214
Title: rest_api: aux_tag creation/updating could set the title to None
Component: REST API
Level: 1
Class: Bug fix
Version: 2.2.0i1
Aux tag creation or updating could result in the title being
set to None or even an empty string. This change prevents that
by setting a min length of 1 for this field.
ID: 15123
Title: Fix lost changes while editing dashboards
Component: Multisite
Level: 1
Class: Bug fix
Version: 2.2.0i1
If you edited a dashboard, e.g. moved a dashlet arround, sometimes these
dashlet was positioned as before after the next change in the dashboard.
This has been fixed.