adalal/gotenberg
1
0
Fork 0
mirror of https://github.com/gotenberg/gotenberg.git synced 2026-08-14 11:22:15 +01:00
Code Issues Packages Projects Releases Wiki Activity

improving documentation

Browse Source
This commit is contained in:
Julien Neuhart
2019-10-02 11:41:35 +02:00
parent 930622ea3c
commit 69e43e40eb
8 changed files with 25 additions and 25 deletions
Download Patch File Download Diff File Expand all files Collapse all files

17
build/docs/content/12-scalability.md
View File

@@ -2,17 +2,18 @@
title: Scalability
---
Google Chrome and unoconv (LibreOffice) are intricate programs.
The API uses under the hood intricate programs.
Gotenberg tries to abstract as much complexity as possible but it can
only do it to a certain extend.
For instance, Google Chrome misbehaves if there are too many concurrent conversions.
That's why for [HTML](#html), [URL](#url) and [Markdown](#markdown) conversions, the API does only 6 conversions in parallel.
The more concurrent requests, the more `504` HTTP codes the API will return.
For instance, [Office](#office) and [Merge](#merge) endpoints will start respectively as many LibreOffice (unoconv) and PDTk
instances are there are requests. The limitation here is the available memory and CPU usage.
On another hand, for [Office](#office) conversions, the API will start as many unoconv (LibreOffice) instances as there are
requests. The limitation here is the available memory.
On another hand, for the [HTML](#html), [URL](#url) and [Markdown](#markdown) endpoints, the API does only 6 conversions in parallel.
Indeed, Google Chrome misbehaves if there are too many concurrent conversions.
**The more concurrent requests, the more `504` HTTP codes the API will return.**
> See our [load testing use case](../loadtesting) for more details about the API behaviour under heavy load.
@@ -20,10 +21,8 @@ requests. The limitation here is the available memory.
### Increase timeout
This strategy is mostly for [HTML](#html), [URL](#url) and [Markdown](#markdown) conversions.
You may increase the conversion timeout. In other words, you accept that a conversion takes more time
if there are more than 6 conversions in parallel.
if the API is under heavy load.
> See [timeout section](#timeout).