Percona Monitoring and Management 2.12.0

Percona Monitoring and Management (PMM) is a free and open-source platform for managing and monitoring MySQL, MongoDB, and PostgreSQL performance.

Release Highlights

• VictoriaMetrics replaces Prometheus and is now the default data source. VictoriaMetrics supports both PUSH (client to server) and PULL metrics collection modes. (Read more.)
• PMM Client can be run as a Docker image.
• The 'Add Instance' page and forms have been redesigned and look much better.

New Features

• PMM-5799: PMM Client now available as docker image in addition to RPM, DEB and .tgz
• PMM-6968: Integrated Alerting: Basic notification channels actions API Create, Read, Update, Delete
• PMM-6842: VictoriaMetrics: Grafana dashboards to monitor VictoriaMetricsDB as replacement for dashboards that used to monitor Prometheus DB
• PMM-6395: Replace Prometheus with VictoriaMetrics in PMM for better performance and additional functionality

Improvements

• PMM-6744: Prevent timeout of low resolution metrics in MySQL instances with many tables (~1000's)
• PMM-6504: MySQL Replication Summary: MySQL Replication Delay graph not factoring in value of intentionally set SQL_Delay thus inflating time displayed
• PMM-6820: pmm-admin status --wait option added to allow for configurable delay in checking status of pmm-agent
• PMM-6710: pmm-admin: Allow user-specified custom 'group' name when adding external services
• PMM-6825: Allow user to specify 'listen address' to pmm-agent otherwise default to 127.0.0.1
• PMM-6793: Improve user experience of 'add remote instance' workflow
• PMM-6759: Enable Kubernetes startup probes to get status of pmm-agent using 'GET HTTP' verb
• PMM-6736: MongoDB Instance Summary dashboard: Ensure colors for ReplSet status matches those in MongoDB ReplSet Summary dashboard for better consistency
• PMM-6730: Node Overview/Summary Cleanup: Remove duplicate service type 'DB Service Connections'
• PMM-6542: PMM Add Instance: Redesign page for more intuitive experience when adding various instance types to monitoring
• PMM-6518: Update default data source name from 'Prometheus' to 'Metrics' to ensure graphs are populated correctly after upgrade to VictoriaMetrics
• PMM-6428: Query Analytics dashboard - Ensure user-selected filter selections are always visible even if they don't appear in top 5 results
• PMM-5020: PMM Add Remote Instance: User can specify 'Table Statistics Limit' for MySQL and AWS RDS MySQL to disable table stat metrics which can have an adverse impact on performance with too many tables

Bugs Fixed

• PMM-6811: MongoDB Cluster Summary: when secondary optime is newer than primary optime, lag incorrectly shows 136 years
• PMM-6650: Custom queries for MySQL 8 fail on 5.x (on update to pmm-agent 2.10) (Thanks to user debug for reporting this issue)
• PMM-6751: PXC/Galera dashboards: Empty service name with MySQL version < 5.6.40
• PMM-5823: PMM Server: Timeout when simultaneously generating and accessing logs via download or API
• PMM-4547: MongoDB dashboard replication lag count incorrect (Thanks to user vvol for reporting this issue)
• PMM-7057: MySQL Instances Overview: Many monitored instances (~250+) gives 'too long query' error
• PMM-6883: Query Analytics: 'Reset All' and 'Show Selected' filters behaving incorrectly
• PMM-6686: Query Analytics: Filters panel blank on Microsoft Edge 44.18362.449.0
• PMM-6007: PMM Server virtual appliance's IP address not shown in OVF console
• PMM-6754: Query Analytics: Bad alignment of percentage values in Filters panel
• PMM-6752: Query Analytics: Time interval not preserved when using filter panel dashboard shortcuts
• PMM-6664: Query Analytics: No horizontal scroll bar on Explain tab
• PMM-6632: Node Summary - Virtual Memory Utilization chart: incorrect formulas
• PMM-6537: MySQL InnoDB Details - Logging - Group Commit Batch Size: giving incorrect description
• PMM-6055: PMM Inventory - Services: 'Service Type' column empty when it should be 'External' for external services

Known Issues

• PMM-7092: Update docker pmm-server 2.11.1 to 2.12.0 results in an unhealthy container.

  Workaround: A folder is not created on container upgrade and will need to be created manually for one of the components. Before starting the new pmm-server 2.12.0, execute:  

  docker exec -ti pmm-server mkdir -p /srv/victoriametrics/data
  docker exec -ti pmm-server chown -R pmm:pmm /srv/victoriametrics/
  docker restart pmm-server