Percona Monitoring and Management 2.31.0

Percona Monitoring and Management (PMM) is an open source database monitoring, management, and observability solution for MySQL, PostgreSQL, and MongoDB.

⚠️ Important
We recommend using the latest version of PMM. This ensures that you have access to the latest PMM features and that your environment runs on the latest version of the underlying components, such as VictoriaMetrics, with all the bug fixes in place.

Release Highlights

Enhaced query building, navigation and Alerting

We have upgraded Grafana to version 9.1 in PMM 2.31.0. With this upgrade, you can get a simplified user interface, enhanced visualizations, and a default unified Alerting experience. For more information, see What’s new in Grafana.

The following features have been introduced with PMM 2.31.0:

• Expandable main menu

  For an enhanced user experience, we have expanded and designed the main menu by moving the dashboard navigation from the PMM Dashboards menu to the main menu for quick access. So you can now browse dashboards like Operating System (OS), MySQL, MongoDB, PostgreSQL, ProxySQL, and HAProxy directly from the navbar.

• Prometheus Query Builder

  We have introduced a new UI query builder using which you can write Prometheus queries effortlessly. Prometheus queries can be complex to grasp, but with this Grafana release, it becomes easier.

• Starred dashboards

  To enhance user experience further, we have introduced Starred, where you can access your favorite dashboards directly from the main menu.

• Explore to dashboard

  You can now create panels and dashboards directly from Explore. You can also create a panel in a new or existing dashboard by clicking Add to dashboard in the Explore toolbar. The generated panel contains all the pane's queries, and the default visualization automatically picked from the current results shown in Explore.

• Improvements to dashboard search

  For a more effective search experience, we have introduced search by title as well as panel. With this, the overall performance has been enhanced as querying and sorting have become much faster.

• Command Palette

  You can now pull up a palette of commands for easier navigation using cmd+k (macOS) or ctrl+k (Linux/Windows).

General availability of Percona Alerting

With this release, we are streamlining the PMM alert setup process with an overhauled, unified alerting system based on Grafana.

All previous Integrated Alerting and Grafana Alerting functionality are now consolidated on the Alerting page. From here, you can configure, create and monitor alerts based on Percona or Grafana templates.

The Alert Rules tab has also evolved into a more efficient interface with an added layer for simplifying complex Grafana rules. You’ll find that the new Percona templated alert option here offers the same functionality available in the old Integrated Alerting, but uses a friendlier interface, powered by Grafana’s advanced alerting capabilities.

As an important and generally useful feature, this new Alerting feature is now enabled by default and ready to use in production!

For more information about Percona Alerting, check out Percona Alerting.

Deprecated Integrated Alerting

The new Percona Alerting feature fully replaces the old Integrated Alerting available in previous PMM versions. The new alerting brings full feature parity with Integrated Alerting, along with additional benefits like Grafana-based alert rules and a unified alerting command center.

If you have been using Integrated Alerting in the previous PMM version, your custom alert rule templates will be automatically migrated to PMM 2.31. After upgrading to this new version, you will find all your alert templates under Alerting > Alert Templates.
These templates are also available for creating new alert rules using the new Percona templated alerts option on the Alert Rules tab.

However, alert rules created with Integrated Alerting are not automatically migrated to Percona Alerting. After upgrading, make sure to manually migrate any custom alert rules that you want to transfer to PMM 2.31 using the Integrated Alerting Migration Script.

Vacuum monitoring dashboard

⚠️ Important
This experimental dashboard is subject to change. It is recommended to use this dashboard for testing purposes only.

A new experimental dashboard has been released to help our users gain timely insights into the autovacuum process in PostgreSQL. This dashboard is designed to help fine-tune the vacuum configuration and to prevent a XID wraparound - both of which directly contribute to improved performance of the database.

This dashboard contains the following:

• Dead tuples - Identifies the number of dead rows in each table even though the rows are physically removed from the table.

• Last time vacuum ran - Tracks the last time a vacuum or autovacuum process successfully ran on each of your tables.

• Number of rows modified since last Analyze - The number of rows changed since the last time ANALYZE ran.

• Manual vacuum events - Tracks the number of times a manual vacuum was run on each table.

• Table disk usage - Tracking the disk space used by each table is crucial as it enables you to gauge expected changes in the query performance over time - but it can also help you detect potential vacuuming-related issues.

For more information, see documentation

New official (GA) deployment method - Podman

We are excited to announce the General Availability (GA) of Podman support for Deploying PMM 2.31.0. We had introduced it in 2.29.0 as a preview feature, but now we are production ready with this feature.

Simplied deployment with Database as a Service (DBaaS)

In our constant endeavor and focus on an enhanced user experience, in PMM 2.31.0, we have simplified the deployment and configuration of DBaaS as follows:

• With PMM 2.31.0, you can easily add a DB cluster from the newly created K8s cluster. All the DB cluster window fields are auto-populated with the values based on the existing K8s cluster.

• For PMM 2.31.0, while accessing DbaaS, if you have an existing Kubernetes cluster configured for DBaaS, you will be automatically redirected to the DB Cluster page. Otherwise, you would be redirected to the Kubernetes Cluster page.

New Features

• PMM-10089: We have upgraded Grafana to 9.1 in PMM 2.31.0.

• PMM-10092: We have expanded and improved the main menu for PMM 2.31.0 based on Grafana 9.1. The main menu contains crucial elements, one of them being starred dashboards.

• PMM-10467: Integrated Alerting: Alert rules created with the Integrated Alerting feature in previous PMM versions can be migrated to the new Percona Alerting using the script that will be provided for the public release of PMM 2.31.

• PMM-10443: Dashboards: We have released a new experimental dashboard for collecting metrics for vacuum monitoring to help our users gain insight into the autovacuum process in PostgreSQL.

Improvements

• PMM-10560: Dashboard: We have improved the MongoDB Collection Details Dashboard. The Collection Activity panel is now collapsed by default, and the graphs are visible only when the--enable-all-collectors parameter is passed in pmm-admin command.

• PMM-10349: DBaaS: For Simplified DBaaS experience, you will be automatically redirected to the DB Cluster page if you have an existing Kubernetes cluster configured for DBaaS. Otherwise, you would be redirected to the Kubernetes Cluster page.

• PMM-10064: Introduced UI changes on the PMM Inventory pages for better readability and user experience.

• PMM-10076: For an improved UI experience, we have enforced the enter button for the date picker, which performs the same action as clicking apply time range.

• PMM-10018: Lowered the default values of CPU and memory for haproxy while deploying a DBaaS Percona XtraDB Cluster (PXC) cluster to avoid wasting resources.

• PMM-9118: Refined the error message thrown when a secure connection cannot be established while adding a monitoring database to PMM, thus making it much more contextual and easy to troubleshoot.

• PMM-10516: Added support for MongoDB physical backups. With this, PMM backup management now supports both logical and physical backups for MongoDB services that are based on Percona Server for MongoDB.

  Percona Backup for MongoDB (PBM) compatibility: MongoDB backups in PMM require Percona Backup for MongoDB version 1.8.x. Make sure not to install the latest BPM version via the percona-release tool, as this automatically installs BPM v.2.0 which is not yet compatible with PMM.

• PMM-10095: While removing the service with pmm-admin, now you do not have to specify the service name for a single service. PMM will automatically look up this service and remove it.

• PMM-10495: By default, Swagger UI attempts to validate specs against swagger.io’s online validator. For PMM 2.31.0, we have disabled it to safeguard your privacy.

• PMM-7806: Upgraded postgres_exporter version used in pmm from 0.8.0 to 0.10.1. With this upgrade, you can access the latest features with all the bug fixes in place.

Bugs Fixed

• PMM-10628: While adding MongoDB or PostgreSQL database via socket, PMM was throwing an error. This has been fixed in PMM 2.31.0.

• PMM-10624: Empty dataname label after PostgreSQL upgraded in PMM.

• PMM-10608: Fixed issues pertaining to the new Home Dashboard, such as high CPU panel unreadable for a higher number of nodes (more than ten nodes), filtering for the environment as well as node name not working correctly, time range not working as expected, some panels on the dashboard not displaying data for selected filters, and so on.

• PMM-10587: Fixed an issue where if you hover over the telemetry icon stacktrace with error appears.

• PMM-10513: Fixed an issue where pmm-admin summary stored PMM password in client/status.json, thus posing a security risk.

• PMM-10065: Fixed an issue encountered while monitoring a large number of nodes (1000) and services. Here, the changing scrape frequency led to data not being captured on the dashboard and throwing a templating error.

• PMM-9973: Fixed an issue where the CPU utilization for Postgres Exporter was on the higher side.

• PMM-10044: Fixed an issue where QAN displayed enormously high values.

• PMM-10218: Fixed an issue for corrupted pprof file on concurrent pprof request.