Percona Monitoring and Management 2.15.0

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

Release Highlights

• PMM 1 vs. 2 Parity

  Disable collectors during adding node/service to monitoring

  With this feature users can disable any collector used by PMM to get metrics. When metrics cannot be collected or are no longer needed, disabling the collector(s) prevents PMM from flooding logs and saves infrastructure resources.

  Our vision for PMM collectors is to provide “stop from collecting” functionality to prevent possible harm to the user environment. This “disable” feature is an initial step towards the ideal functionality. The full and flexible management for “What metrics to collect and in what resolution” is slated for future releases.

  External services monitoring

  Since PMM 1.4.0, users had the ability to monitor external services Percona didn’t currently support (e.g., Redis). This blog article from 2018 nicely described external services monitoring at that time. (At that time Percona was not natively supporting a PostgreSQL monitoring service and so this was listed as an external service. Today, PostgreSQL is natively supported by PMM.)

  Until now, PMM 2.x didn’t support external services monitoring. With this release, any non-natively supported by PMM service will now become supported with external services monitoring. You can see the list of possible exporters to be used in https://prometheus.io/docs/instrumenting/exporters/. Natively-supported services will continue to deliver an expanded set of metrics and insights.

  Provide summary information for systems (pt-*-summary actions)

  With the addition of pt-*-summary in PMM 2, users can now view summary information about services and nodes on PMM’s dashboard. This summary information is in the industry common format of pt-*-summary tools output to simplify portability of this data. This format will also be preserved in the snapshot of the dashboard shared with Percona Support to simplify investigations of issues.

  Note: pt-*-summary includes formats for:

  • pt-mysql-summary
  • pt-mongodb-summary
  • pt-pg-summary
  • pt-summary

• HAProxy support by PMM

  Users are able to add HAProxy Services for monitoring in PMM2. The support level of them in PMM will be the same we have for ProxySQL, so they will be presented in Inventory and on Dashboard. This will allow users who use HAProxy in their HA configuration to also have this component monitored by PMM. In future releases PMM will start use HAProxy by default for the DBaaS feature and will also use this functionality to monitor HAProxy.

• DBaaS Preview improvements (Technical Preview)

  From now you will be able to see the progress of internal steps the system makes when executing some operations with DBaaS. The Progress Bar will not be time-related and will present only steps. The Progress Bar component will also reflect the K8s/Operator-related errors to the user, so in the case of errors, you will have the error text on the UI, and no need to use K8s tools to see the error. With the same UI, you will be able to see the latest logs from K8s so they will have even more information about why the error happened.

  Known Limitations: The progress bar will not provide valuable information for the Delete operation (will be in a later version when we’ll change the API with Operators Team), Operation of DB Cluster Modification will have “strange” behavior and will start changes from non-zero values of steps. (This will be modified after API changes.)

New Features

• PMM-4172, PMM-4306, PMM-5784, PMM-7177: Services and Nodes Summary presentation. Present information about DB's and Node status using pt-mysql-summary, pt-mongodb-summary, pt-pg-summary outputs (in API and on Dashboards).
• PMM-7123: Ability to add External Services via the UI in PMM server.
• PMM-6711: Add external-group flag for pmm-admin inventory commands for simpler work with External services.
• PMM-7405: Check connection response format when adding External Service to monitoring.
• PMM-6797: HAProxy monitoring: Ability to add HAProxy services with pmm-admin [inventory] add [service] haproxy command.
• PMM-7487: HAProxy monitoring: Check connection to HAProxy services when adding them for monitoring.
• PMM-7496: HAProxy monitoring: New HAProxy PXC dashboards.
• PMM-6943: HAProxy monitoring: Show HAProxy type services in PMM Inventory.
• PMM-6924: Integrated Alerting: Show 'breadcrumbs' navigation aid on non-dashboard pages as well as Grafana dashboard pages.
• PMM-7294: Integrated Alerting: Pagination for viewing large numbers of Alert Rules.
• PMM-7417: Security Threat Tool: Show list of all available security checks.
• PMM-7418: Security Threat Tool: Ability to disable specific security checks.
• PMM-7419: DBaaS: Ability to see DB Cluster creation/modification logs.
• PMM-7266: DBaaS: Cluster creation progress bar -- You can now see the progress of DBaaS DB cluster creation. (The progress bar is based on the number of back-end technical steps, not the time required to perform the tasks.)

Improvements

• PMM-4679: Docker: :latest tag for pmm-server and pmm-client images has been moved from v1 latest release to v2 latest release. Note: use of the latest tag is not recommended in production environments, instead use :2 tag.
• PMM-7472: Remove Prometheus data source -- If you were using custom dashboards with a specified data source (not using empty to use default one) you may need to edit your dashboards to use the proper data source. PMM is no longer using Prometheus but uses compatible storage for metrics from VictoriaMetrics. We renamed the data source to be more technology-agnostic.
• PMM-6695: Software update: Grafana 7.1.3 to 7.3.7 (See What's new in Grafana 7.2 and What's new in Grafana 7.3.)
• PMM-7471: Software update: VictoriaMetrics 1.52.0 to 1.53.1 (See VictoriaMetrics 1.53.0 and VictoriaMetrics 1.53.1.)
• PMM-6693: API keys usage -- PMM users can now use API keys (generated in Grafana UI) for interaction with PMM server instead of username/password pairs. The API key should have the same level of access (Admin or Viewer) as is required for username/password pairs.
• PMM-7240: DBaaS: Change from Dashboard to Grafana Page -- We changed the DBaaS page from a Grafana Dashboard to a Grafana Page to be better aligned with the DBaaS enable/disable status and avoid confusion when DBaaS is disabled.
• PMM-7328: Security Threat Tool: Download and run checks when activated, immediately, repeating every 24 hours thereafter (Previously, downloading and running new checks happened every 24 hours but the cycle didn't begin when STT was activated.)
• PMM-7329: Security Threat Tool: Hide check results tab if STT is disabled.
• PMM-7331: Security Threat Tool: Failed checks have 'Read more' links with helpful content.
• PMM-7422: Security Threat Tool: View all active and silenced alerts.
• PMM-7257, PMM-7433: Integrated Alerting: Easier-to-read rule details in Alert Rules list (API and UI presentation).
• PMM-7259: Integrated Alerting: Better UI error reporting for disabled Integrated Alerting. (Hint to users how to enable it.)
• PMM-5533: Better indentation of columns in pmm-admin list output.
• PMM-5888: Improve pmm-admin --help descriptions for external services.

Bugs Fixed

• PMM-5837: pmm-agent reports "Malformed DSN" error when adding PostgreSQL instance with a PMM user password containing = (equals sign) (Thanks to Alexandre Barth for reporting this issue).
• PMM-5969: Removing Services or Nodes with pmm-admin ... --force mode does not stop running agents, VictoriaMetrics continues collecting data from exporters.
• PMM-6685: In low screen resolutions Services submenu wraps, becomes obscured, and can't be accessed.
• PMM-6681: Not all PMM admin users can download diagnostic logs, only those with Grafana admin rights.
• PMM-7227: Table stats metrics not being collected in instances with millions of tables.
• PMM-7426: vmagent continually restarts, blocking comms between pmm-agent & pmm-managed -- Users running multiple services on the same PMM agent in 'push' mode could face this issue when restarting the agent after bulk-adding services.
• PMM-6636: Dashboards: MySQL Replication Summary: 'Binlog Size', 'Binlog Data Written Hourly', 'Node' not being charted when the instance is RDS.
• PMM-7325: Dashboards: MySQL User Details: user labels unreadable with high number (>20) of users (Thanks to Andrei Fedorov for reporting this issue).
• PMM-7416: Dashboards: PostgreSQL Instance Summary: Some panels (e.g. Tuple) not using selected database.
• PMM-7235: Integrated Alerting: Filtered out alerts are shown in the UI as firing.
• PMM-7324: Integrated Alerting: Add Pager Duty Notification Channel: after user pastes copied key Add button is not enabled.
• PMM-7346: Integrated Alerting: It is possible to create Alert Rule with negative duration time.
• PMM-7366: Integrated Alerting: Entities (e.g. templates, channels, rules) are in inconsistent states.
• PMM-7467: Integrated Alerting: < (less-than symbol) wrongly interpreted by Alert templates (as <).
• PMM-7591: Integrated Alerting: User can not receive notifications on email after password update.
• PMM-7343: Security Threat Tool: Check results show previously failed checks after STT re-enabled.
• PMM-7250: DBaaS: Confusing error "Cannot get PSMDB/PXC cluster" appears after removing DB cluster.
• PMM-7193: DBaaS: Number of Nodes can be set as float.
• PMM-7349: DBaaS: Host and Password occasionally disappearing from Connection column.