Skip to content

v2.34.0

Compare
Choose a tag to compare
@github-actions github-actions released this 31 Jan 14:49
· 1248 commits to main since this release
v2.34.0

Percona Monitoring and Management 2.34.0

Release date: January 30, 2023
Installation: Installing Percona Monitoring and Management

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

Release Highlights

Various UX enhancements to Backup Management

With this release, we are updating the Backup and Restore functionality with key usability enhancements, including:

  • Filtering on services at backup creation time. This makes it considerably easier to find the right service for a backup when working with multiple servers.
  • Automatic redirection to the Restores page after initiating a restore. From here you can immediately monitor the restore process, and see the restore history for your PMM instance.
  • Added information about the type of storage for the backup artifacts on the All backups page. This helps you get a more comprehensive overview of your current backups.
  • Design and usability improvements on the Create Backup and All Backups page to increase consistency with other PMM UI controls.

DBaaS migration

Starting with PMM 2.34.0, PMM installs and manages operators using Operator Lifecycle Manager (OLM) and database clusters using dbaas-operator. Nonetheless, these changes are backward incompatible, and you must perform these additional steps:

  1. Force unregister the Kubernetes cluster from PMM (select Ignore errors checkbox) and register it back. PMM will install OLM and dbaas-catalog to manage the lifecycle of PXC, PSMDB, and DBaaS operators.
  2. Create a database cluster for each cluster you have created in PMM. You can use the script to do it automatically during the maintenance window. Migration scripts use the default configuration to connect to Kubernetes clusters and run migrations against them.

New components

We have a newly added component, VMProxy starting with PMM 2.34.0. VMProxy is a stateless reverse proxy that proxies requests to VictoriaMetrics and optionally adds extra_filters query based on the provided configuration.

New Features

PMM-11031 - To enhance the DBaaS experience, the UI has been extended to include the following:

  • Network and Security - DBaaS users can set a source range from which they can connect to their DB clusters with the Expose field. You can modify the exposure level using the Source Range field. Additionally, the Source Range field allows you to choose the network from which you can connect.
  • Configuring the database engine - Copy-paste the config string into the Configuration text box to configure the database engine.
  • Configuring storage classes - You can choose from the available classes.

Improvements

  • PMM-11024 - [DBaaS]: Move from dbaas-controller to dbaas-operator
  • PMM-10873 - [DBaaS]: Create DB Cluster: We have modified the UI to include a separate window rather than a popup so that creating a DB cluster is easier.
  • PMM-11132 - [DBaaS]: During K8s cluster registration, a new state called Provisioning is added. Provisioning is the default state, and once that is completed, the state will change to Active.
  • PMM-11134 - [DBaaS]: Edit DB Cluster: We have modified the UI to include a separate window rather than a popup so that editing a DB cluster is easier.
  • PMM-9262 - [Backups]: After starting a backup restore from the All Backups page, PMM now automatically redirects you to the Restores tab, where you can check the status of your new restore.
  • PMM-11086 - [Backups]: Improved design for the On Demand and Schedule Backup cards on the Create Backup page. In addition, the Created column on the Restores tab has been replaced with Started at and Ended columns to provide more insight into the restore process.
  • PMM-10924 - [Backups]: The Location column on the All Backups page has been enhanced to include information about the storage type used for the backup: Local or S3-compatible.
  • PMM-11171 - [Backups]: The Service name field on the Create Backup page is now searchable. This makes it easier to find the right service for a new backup when working with multiple servers. Just paste the name of the server you're looking for, or start typing to see the list of matching servers.
  • PMM-9079 - [Backups]: Added Target Service column under Backup > Restores to show the service to which a backup was restored.
  • PMM-10773 - [Alerting]: Added option to unsilence alerts from the Actions menu under Alerting > Fired Alerts.
  • PMM-11230 - [Alerting]: The Fired alerts page under Alerting now auto-refreshes more frequently to ensure that any new alerts are displayed in real-time.
  • PMM-11255 - [Alerting]: PMM no longer transfers the name of the current alert rule when switching from Percona templated alert to Grafana managed alert.
  • PMM-11090 - The Logs feature has been moved from the dbaas-controller to PMM, providing users with the ability to see the events for each pod and the logs for every container properly populated within PMM.
  • PMM-11179 - Starting with PMM 2.34.0, we support push metrics mode while updating labels.
  • PMM-11189 - Starting With PMM 2.34.0, we have implemented vmproxy integration and frontend API changes.
  • PMM-10705 - [DBaaS]: As of PMM 2.34.0, Operator Lifecycle Management (OLM) is used for handling the installation and updating of operators for PMM in an effective, automated, and scalable manner.
  • PMM-11069 - [Backups]: PMM no longer allows creating backup tasks with a name that already exists. After upgrading to PMM 2.34, any backup tasks with duplicate names are automatically renamed to apply the new naming constraints.

Bugs Fixed

  • PMM-7350 - [DBaaS]: Fixed an issue where DB Cluster did not restart all PXC servers and proxy SQL servers on restarting.
  • PMM-11333 - [DBaaS]: Fixed an issue where the API key was missing on K8s cluster registration after moving to Operator Lifecycle Management (OLM).
  • PMM-9505 - [DBaaS]: The options to update, restart, and view logs were enabled for paused clusters. This issue has now been resolved by disabling these options.
  • PMM-11221 - [DBaaS]: Fixed an issue where the page did not load when DBaaS was disabled.
  • PMM-7958 - During Postgres monitoring, databases were not getting deleted. This issue has been resolved now.
  • PMM-9977 - [Backups]: After deleting an expanded row from the Backup > All Backups page, PMM would automatically expand the next row in the table. This issue is now fixed and PMM refreshes the list of backups artifacts without automatically expanding any rows.
  • PMM-10942 - [Backups] - The scheduler for PITR backups now use clearer text labels to better define the Start time and the Frequency for the backups being created.
  • PMM-10853 - [Alerting]: Removed Save button from the Create alert rule page as this was duplicating the functionality of the Save and exit button.
  • PMM-10871 - [Alerting]: Fixed broken Percona Alerting link on the Settings > Alertmanager integration page. The link now correctly redirects to the Alerting page.
  • PMM-11049 - Fixed a typo that was present in enhanced metrics scraper for rds_exporter.
  • PMM-11137 - An issue with the MongoDB exporter not showing the version for MongoDB instance has been resolved.
  • PMM-11210 - Whenever a cluster was unavailable or deleted, the K8s cluster list was empty. The issue has been resolved now.
  • PMM-11239 - Despite removing pmm-server packages before upgrading PMM server, the packages were still present during upgrade. The issue has been resolved now.
  • PMM-11372 - The user interface was broken after PMM was upgraded. This issue has been resolved now.

Known issues

  • Starting with PMM 2.34.0, the recommended version for running MySQL cluster is 8.0.27. It's impossible to create MySQL clusters using a version older than 8.0.27. All previously created clusters are not affected and will be migrated without any issues.
  • PMM-11474: If more than two Kubernetes clusters are registered with PMM, and database cluster credentials are requested, an internal server error may occur.