Maintenance and update policy in PostgreSQL Flex
Last updated on
To ensure the security, stability, and performance of your database instances, updates for the underlying operating system and PostgreSQL are provided on a regular basis. We announce maintenance windows at least two weeks in advance. Depending on the type of update, scheduling and customer control are handled as follows:
Minor updates and OS patches
Section titled “Minor updates and OS patches”- Process: New PostgreSQL minor versions (e.g., from
18.6to18.7) and routine operating system patches are automatically applied to your instances after passing internal validation and testing. - Scheduling: Installations take place exclusively within an announced maintenance window.
Critical security patches
Section titled “Critical security patches”- Process: To protect your data against critical vulnerabilities in the infrastructure or in PostgreSQL, security patches are deployed immediately after release and verification.
- Scheduling: Due to the urgency of security fixes, these patches are applied as promptly as possible and may occur outside regular maintenance windows without extended advance notice.
Major upgrades
Section titled “Major upgrades”- Process: Upgrades to a new PostgreSQL major version (e.g., from PostgreSQL 17 to 18) are not performed automatically. You remain in full control and decide when to initiate a major version upgrade.
- End-of-Life (EOL): When a PostgreSQL major version approaches its official end-of-support date, you will be notified in advance so that you can perform an upgrade. After a defined grace period expires, any remaining instances on the unsupported version will undergo a mandatory automatic upgrade (forced upgrade) to maintain service security and compliance.
Impact on service continuity (downtime)
Section titled “Impact on service continuity (downtime)”- Single-node instances: Applying OS patches or database updates requires a service restart, resulting in a brief temporary downtime typically lasting a few minutes.
- High-Availability setups (HA / Multi-node): Updates are performed in a rolling manner across the nodes, which may cause brief interruptions.