[Release] Checkmk beta release 2.1.0b8

Dear friends of Checkmk,

the new beta release 2.1.0b8 of Checkmk is ready for download.

This test release ships with 7 changes affecting all editions of Checkmk,
1 Enterprise Edition specific changes and 1 Managed Services Edition specific
changes.

Additionally, we wanted to let you know that we are also preparing our user
guide for the new version.
Here you can find some articles that are relevant for new features of 2.1:

New agents: Monitoring Linux - The new agent for Linux in detail & Monitoring Windows - The new agent for Windows in detail
Kubernetes: Monitoring Kubernetes
Influx: Sending metrics to InfluxDB and Graphite - via InfluxDB v2 API or Carbon plaintext protocol
Grafana: Integrating Checkmk in Grafana
Datadog: Integrating Datadog
Two-factor authentication: Users, roles and permissions - Authorization and user configuration

Changes in all Checkmk Editions:

Checks & agents:

  • 13857 Windows agent uses perflib to obtain wmi_cpuload data
  • 14121 KUBE: Import annotations as host labels
  • 14084 Deprecate old HW/SW inventory plugin API
    NOTE: Please refer to the migration notes!
  • 14123 FIX: KUBE: Change Service name ‘Cluster Collector’
    NOTE: Please refer to the migration notes!

Core & setup:

  • 13931 FIX: wrong labels not raising errors in the REST API

Setup:

  • 14102 FIX: Fix possible ValueError on activation of remote sites
  • 13748 FIX: Support Diagnostics: Missing performance report

Changes in the Checkmk Enterprise Edition:

User interface:

  • 14104 FIX: Fix error on adding views to reports

Changes in the Checkmk Managed Services Edition:

Setup:

  • 14103 FIX: Fix synchronisation of password store entries to customer sites

You can download Checkmk from our download page:

Please mail bug reports and qualified feedback to feedback@checkmk.com.
We greatly thank you for using Checkmk and wish you a successful monitoring,

Your Checkmk Team

After installing the b8 on some smaller systems i saw i problem that can be really big.
With the separation of “Check_MK” and “Check_MK Agent” service we lost the complete information about the agent/snmp execution time.
@moritz this execution time information is very crucial for bigger systems to find the problem hosts.
In the systems i tested all the hosts have only 10 to 50ms execution time and this is not true.
The execution time of the agent call and the SNMP response time must be somewhere visible.
I had bigger systems where a hand full of systems killed the performance of the complete monitoring system with very long response times.

1 Like

I cannot reproduce this behavior. In fresh sites on my localhost I see timings in the expected range ~4 seconds. Am I missing something here?

It looks like that it only affects the RAW edition. That i only saw after some more tests with other small sites.

I can confirm that. It’s been introduced in b3. I’ll fix it, but I am afraid the issue will remain in b9.

1 Like