mirror of
https://github.com/andwati/zitadel.git
synced 2026-10-18 05:52:51 +00:00
Wim Van LaerandGitHub
9b064d8b74
# Which Problems Are Solved - Typos - Punctuation - Markdown table formatting - ... (A bunch of issues my editor gave) # How the Problems Are Solved # Additional Changes # Additional Context
38 lines
1.3 KiB
Plaintext
38 lines
1.3 KiB
Plaintext
---
|
|
title: Technical Advisory 10004
|
|
---
|
|
|
|
## Date and Version
|
|
|
|
Version: 2.39.0
|
|
|
|
Date: 2023-10-14
|
|
|
|
## Description
|
|
|
|
Due to storage optimizations ZITADEL changes the behavior of sequences.
|
|
This change improves command (create, update, delete) performance of ZITADEL.
|
|
|
|
Sequences are no longer unique inside an instance.
|
|
From now on sequences are counting up per aggregate id.
|
|
For example sequences of newly created users begin at 1.
|
|
Existing sequences remain untouched.
|
|
|
|
## Statement
|
|
|
|
This change is tracked in the following PR: [new eventstore framework](https://github.com/zitadel/zitadel/issues/5358).
|
|
As soon as the release version is published, we will include the version here.
|
|
|
|
## Mitigation
|
|
|
|
If you use the ListEvents API to scrape events use the creation date instead of the sequence.
|
|
If you use sequences on a list of objects it's no longer guaranteed to have unique sequences across the list.
|
|
Therefore, it's recommended to use the change data of the objects instead.
|
|
|
|
## Impact
|
|
|
|
Once this update has been released and deployed, sequences are no longer unique inside an instance.
|
|
ZITADEL will increase parallel write capabilities, because there is no global sequence to track anymore.
|
|
Editor service does not respond the different services of ZITADEL anymore, it returns zitadel.
|
|
As we are switching to resource based API's there is no need for this field anymore.
|