Percona Monitoring and Management 2.37.0

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

Release Highlights

Improved Inventory view

This release of PMM starts the series of enhancements that will improve the way you work with Services and Nodes in PMM.
With this first set of changes, we’re revamping the Inventory page to give you more context on your inventory objects, and more actionable information for exploring and fixing possible issues.

Here are all the specific changes:

• Improved design of the Inventory page for a clearer presentation of all the objects registered by PMM.
• Services now show their database status, which is based on the monitoring metrics. This enables you to quickly check the status of the databases when you are working with the inventory.
• New Monitoring column showing the live status for the monitoring Agents assigned to a Service. This enables you to quickly spot any possible issues with the Agents.
• Quick access to the Service Summary Dashboard and to QueryAnalytics results so that you can easily explore important related information.
• New Details section grouping together all the relevant parameters and labels for a Service. To keep things clean, this section is collapsed by default, so that you can view extra information only when you need it.

Here's what's coming:

Building up on these changes, subsequent releases will further improve PMM Inventory with:

• More useful features to filter and better connect Services and Nodes.
• Services organized in clusters so that you can capture the bigger picture, instead of working at of individual Service-level.

New Experimental Dashboards for Kubernetes Monitoring

⚠️ Important
These experimental dashboards are subject to change. It is recommended to use these dashboards for testing purposes only.

We are pleased to announce the release of PMM 2.37, which introduces three new experimental dashboards:

• Kubernetes Cluster Summary
• Kubernetes Pods Status
• Kubernetes Volumes

These dashboards are designed to provide valuable insights into the status and performance of your Kubernetes cluster, pods, and volumes, helping you to identify and troubleshoot issues quickly and easily.

We welcome your feedback as we continue to enhance PMM with these new dashboards.

Experimental Kubernetes Cluster Summary Dashboard

Kubernetes Cluster Summary provides a comprehensive overview of your Kubernetes cluster, including:

• Components
• Node
• Pod
• PVC status
• CPU
• Memory overview, and more.

This dashboard displays all workloads running in the cluster, enabling you to take action and optimize its performance.

Experimental Kubernetes Pods Status Dashboard

Kubernetes Pods Status dashboard provides detailed information about the state and performance of your pods, including CPU, Memory, and Network metrics.

This dashboard can help you quickly pinpoint any issues affecting your pods and ensure they continue to operate smoothly.

Experimental Kubernetes Volume Dashboard

Kubernetes Volumes dashboard provides insights into your Kubernetes volumes, including capacity and usage, in real time. With this dashboard, you can easily monitor the performance and usage of your volumes and take proactive measures to ensure their performance.

Refer to the documentation to learn more about these new experimental dashboards and how to use them.

Here are the steps to create a new folder and move all experimental dashboards to the new folder for quick access and internal use:

📝 Note
You should have at least an Editor role to create a new folder and move all experimental dashboards.

1. Navigate to the Main menu and hover on the {{icon.dashboards}} Dashboards icon.
2. Click New folder.
3. Provide a name for your folder, and then select Create.
4. Navigate to {{icon.dashboards}} Dashboards from the Main menu and click Browse.
5. Select the dashboard that you want to move and click Move.
6. On the Choose Dashboard dialogue box, from the dropdown under Folder option, choose the folder where you want to move the dashboard.
7. To apply your changes, select Save Dashboard.

Components upgrade

VictoriaMetrics has been upgraded to version [1.89.1].

New Features

• PMM-10913 - Starting with PMM 2.37.0, you can use an external PostgreSQL server as data storage for PMM. Thus, even if PMM fails, your data will be stored in an external source, reducing the risk associated with data loss. Furthermore, clustered PostgreSQL instances also enhance performance.

• PMM-11281 - We have written a query for the postgres exporter to deal with lock conflicts in databases that block several transactions, thus impacting performance. With this query, you can find the PID of the blocking session and how many sessions it blocked. The proactive approach enables customers to address non-responsive databases quickly by understanding the underlying causes.

• PMM-11384 and PMM-11834 - PMM 2.37.0 now supports PSMDB operator versions 1.13 and 1.14.

• PMM-11438 - Starting with PMM 2.37.0, PMM administrators can disable the internal PostgreSQL server, which helps to fine-tune the server.

• PMM-11439 - Starting with PMM 2.37.0, a PMM administrator can now view metrics from an external PostgreSQL server.

Improvements

• PMM-11412 - We have enhanced our dashboards to be more user-oriented, displaying the DB cluster's status along with its components, such as metrics of stateful, pods, and so on. Also, you can apply filters based on DB name or DB type.
• PMM-11832 - When hovering over the Backup option on the main menu, PMM now displays the subcategories for easier navigation.
• PMM-11522 - With the newly added support for Clickhouse as a datasource for Advisors, Advisors can now use data produced by Query Analytics (QAN) to run checks on databases.
• PMM-11544 - Improved functionality for EXPLAIN query in QAN.
• PMM-11926 - Updated the format of the Upgrade to a Premium plan URL on the Advisor Insights page.
• PMM-11078 - Following the replacement of Integrated Alerting with Percona Alerting in PMM 2.311, we have now phased out all the APIs related to this deprecated feature.

Bugs Fixed

• PMM-8690 - Fixed an issue where the Replica set lag displayed incorrect values on the MongoDB ReplSet Summary dashboard for an ARBITER node.
• PMM-9775 - QAN: PMM failed to display the set timezone for QAN despite it being set to UTC.
• PMM-10687 - DBaaS: Fixed an issue where DB cluster creation failed when using an unreleased version of pmm-server.
• PMM-11121 - DBaaS: Fixed an issue where the list of DB clusters did not load if one of the K8s clusters was unresponsive.
• PMM-11226 - Fixed an issue where different operator versions and names were displayed for each K8s cluster.
• PMM-11313 - Rows Fetched/Read and Users by Rows Read dashboards are not working in the MySQL User Details dashboard
• PMM-11555 - For thousands of client connections for ProxySQL, the graph for Active Frontend connections did not render due to sluggish performance. Furthermore, this could affect the rendering and usage of the ProxySQL Instance summary dashboard. The issue has been resolved now.
• PMM-11829 - PMM Agent does not work with PG 12 and PGSM 2.0. This issue has been resolved now.
• PMM-11844 - Fixed an issue where a user failed to add PSMDB with pmm-admin using native LDAP authentication and TLS.
• PMM-11862 - Fixed inconsistent naming of the Advisors feature across PMM UI.
• PMM-11875 - Fixed an issue where the users could not select and compare multiple nodes on the Summary dashboard.
• PMM-11904 - Index Size and Concurrent Inserts panels were broken on the Insight / VictoriaMetrics dashboard after updating to version 1.89.1.
• PMM-10795 - Node summary was not visible under MongoDB ReplSet Summary when the node name contained a dot (.) in the name.
• PMM-11465 - Fixed problem with empty output for EXPLAIN in QAN.
• PMM-11729 - Fixed syntax issues with placeholders for EXPLAIN in QAN.
• PMM-11849 - Fixed issue that affected Advisors, where Actions and Jobs would always be executed on the primary node in a MongoDB cluster.
• PMM-11934 - Product tour is now correctly displaying the Advisors information.