Skip to content

Percona Server for MySQL

Description

MySQL is the leading database management system (DBMS) for highly reliable, web-based applications; it is supported by numerous shop systems and is recognized globally as the most popular open-source database. By using Percona MySQL, we have opted for a database server that is fully compatible with MySQL yet enhanced to deliver superior performance and scalability. We handle the installation and operation of your database servers, ensuring optimal configuration in coordination with connected systems. Our monitoring service keeps a constant watch on the system, proactively alerting us to issues such as overload or potential data loss.

We optimize cache sizes to deliver that extra boost in speed. Databases and their integration can be complex, depending on your specific requirements. If you are unsure which of our options best suits your application, please feel free to contact our consulting team. Our experts will assist you in meeting your performance, availability, and resilience requirements. Full database backups are performed daily and separately from the Managed Cloud Instance backups. Our solution ensures consistent database backups; if a secondary instance is present, the backup is generated from it, ensuring the primary instance's performance remains unaffected. Retention periods are 7 days, 4 weeks, and 3 months. Our goals for this service:

  • High-performance, reliable database services
  • High configuration flexibility
  • Data hosting in ISO 27001-certified data centers

What you can expect from our service

  • Consistent daily database backups
  • Scalable architecture
  • Monitoring of all relevant health and performance metrics
  • Up to 100 GB of storage for your database
  • Enabled slow query log

Optional service features

  • Additional database server instances
  • More than 100 GB of secondary storage
  • Backup from secondary (if available)
  • Auto-failover (only with ProxySQL or HAProxy)
  • Performance tuning

Service requests

  • Service restart
  • Scaling performance
  • Simple configuration changes
  • Service stop or uninstallation
  • User setup
  • Database setup

Standard Configurations

We offer two distinct standard hosting solutions that differ in performance and availability. Customers are responsible for importing database content—such as data, schemas, and indices. If these options do not meet your requirements, or if you would like assistance with setup, please contact our consulting team to develop a solution tailored to your specific needs.

Single-Server Setup

This option is particularly suitable for development and testing systems, as it offers excellent value for money. A single database service is installed and operated as the "Primary," handling both read and write operations. While adding a "Secondary" server later is generally straightforward, doing so results in a service interruption. Backups are performed while the system is running, which may impact service performance.

Replication Setup (Our Best Practice)

For production environments, we recommend a replication setup that separates read and write operations. All write operations are directed to the Primary, while read operations are handled by the Secondary (provided your client application supports this). This prevents read operations from being blocked and results in higher performance. Read performance can be scaled without interruption by adding additional Secondaries; if the Primary fails, we ensure that a Secondary takes over the Primary's duties. As a preventive measure, we alway configures Secondaries identically unless expressly agreed otherwise (this is a requirement for the Premium SLA). All additional Secondaries synchronize automatically with the Primary. We continuously monitors the time lag (delta) between changes on the Primary and the Secondaries, taking action if a specific threshold is exceeded. Database backups are generated by the Secondaries, reducing the load on the Primary. We offer an optional Premium SLA for this configuration. Trending Our Trending tool gives you access to historical data such as

  • number of connections,
  • buffer pool sizes and usage,
  • query types,
  • queries per second by type, allowing you to monitor and optimize your client application's use of the database servers.

External Access

We do not grant access to the database services from the public internet. Please contact our consulting team if you require this.

Optional Features/Configurations

Distributing the DB Setup Across Multiple Availability Zones

Except for the single-server setup, any configuration variant can be distributed across multiple availability zones. This ensures the service remains available—or can be rapidly restored—even if a complete data center failure occurs.

XtraDB Cluster

This multi-primary configuration, comprising at least three database nodes, ensures full database availability even if two nodes fail. Once a node restarts, it synchronizes itself automatically. To ensure the application can still connect to the database if the specific node it was using fails, we always deploy XtraDB Cluster in combination with ProxySQL, which handles automatic failover.

ProxySQL

With ProxySQL, applications benefit from distributed read requests across multiple database instances. This significantly increases data read speeds while simultaneously reducing the load on your database servers. ProxySQL runs on every application server, ensuring redundancy from the second server onwards. ProxySQL’s connection pooling improves efficiency by reusing existing database connections. Optionally, enabling query caching can further boost read performance by allowing you to define cache validity periods for individual queries; caching takes place locally on the application server. Alternatively, HAProxy can be used, though it does not feature a query cache.

Managed Cloud Instance Sizes

Upon request, we would be happy to provide a quote for larger or smaller Managed Cloud Instance sizes tailored to your performance requirements. High-performance storage options are also available.

FAQ

Can the primary and secondary nodes be located in different availability zones?

Yes, this is possible.

What happens if the database disk space runs low?

Our alerting system triggers an alarm as the disk space limit approaches. We attempt to temporarily increase the storage capacity in the short term; however, it remains your responsibility to stay within the disk storage limits. Please contact our Service Management team to receive a custom quote for servers requiring more than 100 GB of permanent storage.

Can the primary and secondary nodes fail simultaneously?

We distribute the primary and secondary nodes across different hardware to achieve greater component redundancy.

Is there a performance drop during database backups?

The specific impact depends on the chosen setup. In a replication setup, only the performance of the secondary node is affected.

Important Notes

We offer Percona Server for MySQL as a managed service. You can find further information and details about our managed services on our website.