From 7f6b1ab9cb2154af1148540724cc6bda2092cf4f Mon Sep 17 00:00:00 2001
From: Martin Weise <martin.weise@tuwien.ac.at>
Date: Mon, 17 Feb 2025 15:03:18 +0100
Subject: [PATCH] Document

Signed-off-by: Martin Weise <martin.weise@tuwien.ac.at>
---
 .docs/api/data-db.md     | 8 ++++++++
 .docs/api/metadata-db.md | 8 ++++++++
 2 files changed, 16 insertions(+)

diff --git a/.docs/api/data-db.md b/.docs/api/data-db.md
index caeee6d171..f84ab8aad1 100644
--- a/.docs/api/data-db.md
+++ b/.docs/api/data-db.md
@@ -22,6 +22,14 @@ Any number of MariaDB ata databases can be integrated into DBRepo, even non-empt
 registered in the Metadata Database to be visible in the [User Interface](../ui) and usable from e.g. the Python
 Library.
 
+## Configuration
+
+By default, the Data Database is configured as a cluster of three nodes where each node has a maximum of 2048 MiB RAM
+available. As recommended by
+[MariaDB](https://mariadb.com/kb/en/mariadb-memory-allocation/#allocating-ram-for-mariadb-the-short-answer), we set
+`innodb_buffer_pool_size=1430M` (70% of the available RAM). If you have more RAM available, you should set the variable
+accordingly to improve the performance.
+
 ## Data
 
 The procedures requires the in parameter of the `hash_table` stored procedure to have the same collation as the
diff --git a/.docs/api/metadata-db.md b/.docs/api/metadata-db.md
index 05d164cf3c..a3f3f2b8b7 100644
--- a/.docs/api/metadata-db.md
+++ b/.docs/api/metadata-db.md
@@ -34,6 +34,14 @@ services:
     ...
 ```
 
+## Configuration
+
+By default, the Metadata Database is configured as a cluster of three nodes where each node has a maximum of 2048 MiB
+RAM available. As recommended by
+[MariaDB](https://mariadb.com/kb/en/mariadb-memory-allocation/#allocating-ram-for-mariadb-the-short-answer), we set
+`innodb_buffer_pool_size=1430M` (70% of the available RAM). If you have more RAM available, you should set the variable
+accordingly to improve the performance.
+
 ## Image
 
 :octicons-tag-16:{ title="Minimum version" } 1.4.4
-- 
GitLab