After the 13 December 2026 release, Ans will gradually start generating new identifiers using UUIDv7 instead of UUIDv4.
UUIDv7 is time-ordered, which makes identifiers more efficient to store and index in databases. This can improve database performance, particularly for tables containing large numbers of records, while retaining the uniqueness and global usability expected from UUIDs.
What does this mean?
- Existing identifiers will not change.
- New records may receive UUIDv7 identifiers after the release.
- Ans will therefore use a combination of UUIDv4 and UUIDv7 identifiers during the transition.
- Identifiers returned by the API will remain UUIDs and can continue to be handled by standard UUID libraries.
- Integrations should treat identifiers as opaque values and must not assume that all identifiers use UUIDv4.
- Integrations should not rely on identifiers being generated in a particular order. Although UUIDv7 identifiers are broadly sortable, they should not replace explicit timestamps or ordering fields.
Why are we making this change?
UUIDv7 provides better database locality because newly generated identifiers are ordered by creation time. This reduces index fragmentation and improves the efficiency of inserts and queries compared with randomly generated UUIDv4 identifiers.
Existing identifiers will remain unchanged to ensure that references stored in customer systems, API integrations, exports, and historical data continue to work.
What do you need to do?
Integrations that validate the UUID version 4 explicitly, use regular expressions restricted to UUIDv4, or make assumptions about identifier ordering may need to be updated. No changes are required for integrations that already parse and store UUIDs according to the standard UUID format.
Comments
0 comments
Article is closed for comments.