mirror of
https://gitlab.opencode.de/bmi/opendesk/deployment/opendesk.git
synced 2025-12-07 16:01:37 +01:00
50 lines
2.0 KiB
Markdown
50 lines
2.0 KiB
Markdown
<!--
|
|
SPDX-FileCopyrightText: 2025 Zentrum für Digitale Souveränität der Öffentlichen Verwaltung (ZenDiS) GmbH
|
|
SPDX-License-Identifier: Apache-2.0
|
|
-->
|
|
|
|
<h1>Size Profiles</h1>
|
|
|
|
This repository contains base profiles for the following maximum user counts:
|
|
|
|
- 500
|
|
- 5,000 (5k)
|
|
- 50,000 (50k)
|
|
- 100,000 (100k)
|
|
|
|
# Assumptions
|
|
|
|
These profiles were created with some assumptions in mind:
|
|
|
|
- Regular load without application-specific emphasis
|
|
- 10% concurrent user load at max on individual applications
|
|
|
|
As user profiles can vary significantly between different instances, you may want to adjust these values.
|
|
For information on adjusting values, refer to the [scaling](./scaling.md) documentation for openDesk.
|
|
|
|
# Hardware Requirements
|
|
|
|
The following table provides an overview of CPU and memory requirements, including [external services](./external-services.md).
|
|
The available memory in the cluster should be between the request and limit values. Actual memory utilization varies depending on usage patterns but will unlikely consume the complete setup limit.
|
|
|
|
| **Profile** | CPUs | Memory (Gi), request | Memory (Gi), limit |
|
|
|-------------|------|----------------------|--------------------|
|
|
| **500** | >25 | 32 | 141 |
|
|
| **5k** | >54 | 86 | 551 |
|
|
| **50k** | >206 | 422 | 2058 |
|
|
| **100k** | >336 | 711 | 3418 |
|
|
|
|
# Location
|
|
|
|
`/helmfile/environments/default-size-<max-user-profile>-overrides/`
|
|
|
|
# Deployment
|
|
|
|
Set the environment variable `SIZE_PROFILE` to `500`, `5k`, `50k`, or `100k` when using helmfile to use one profile.
|
|
|
|
# Notes
|
|
|
|
The size profiles contain resource definitions for [external services](./external-services.md) that can be used as guidance for production-grade service requirements.
|
|
|
|
Large and very large instances (50k and 100k) also require more allowed concurrent connections (`max_connections`) on the databases (especially PostgreSQL) than the default value (100).
|