• [28352] Improved the Messages panel of the RabbitMQ Grafana dashboard to display absolute values instead of rates.

  • [28066] Fixed the issue with the Host API Status graph being unavailable in the Cinder Grafana dashboard.

  • [26450] Fixed the Apache meta for Telegraf to use the parameters from server.mods.status instead of apache:server:bind.

  • [28123] Fixed the issue with the absent() function causing malfunction of the Ceph Grafana dashboards in case if one of the Prometheus servers had no data for a particular period of time.

  • [27250] Added support for the containerd log format to fix the issue with the inability to parse the Kubernetes container logs.

  • [27142] Fixed the discrepancy in RAM usage data between the Horizon web UI and the Nova - utilization dashboard in Grafana.

  • [26918] Fixed the issue with the false negative http_response_status metric for the Aodh URL by adding support for the HTTP response code 200 for the Aodh checks in OpenStack version Pike and newer.

  • [27982] Fixed the issue with the Apache Grafana dashboard incorrectly displaying a high percentage (thousands of percents) in the CPU Load panel for the ctl nodes.

  • [27474] Removed the non-valuable ContrailFlow* alerts to prevent the false positive raising of such alerts.

  • [27342] Adjusted the NginxServiceDown alert by adding the for: 1m variable to prevent raise of false positive alerts for the NGINX service being down.

  • [27298] Fixed the issue with the inability to resolve the PacketsDroppedByCpuMajor alert in a time frame of less than 24 hours.

  • [26842] Updated the monitoring interval in Telegraf to 40 seconds for Ceph Jewel to prevent timeouts in Telegraf while gathering the data.

  • [24810] Improved regexp for the HDD metrics to prevent generation of false positive for HDD errors.

  • [26116] Added the Fluentd label for Telegraf to fix the issues with processing severity of the Telegraf logs.