[Backport 8.x] Add doc about timeout best practices (#2386)

Co-authored-by: Josh Mock <joshua.mock@elastic.co>
This commit is contained in:
github-actions[bot]
2024-10-21 11:04:55 -05:00
committed by GitHub
parent fceebae8ae
commit e279b3ebfa
3 changed files with 13 additions and 1 deletions

View File

@ -2,10 +2,11 @@
== Configuration
The client is designed to be easily configured for your needs. In the following
The client is designed to be easily configured for your needs. In the following
section, you can see the possible options that you can use to configure it.
* <<basic-config>>
* <<advanced-config>>
* <<timeout-best-practices>>
* <<child>>
* <<client-testing>>

View File

@ -21,3 +21,4 @@ include::reference.asciidoc[]
include::examples/index.asciidoc[]
include::helpers.asciidoc[]
include::redirects.asciidoc[]
include::timeout-best-practices.asciidoc[]

View File

@ -0,0 +1,10 @@
[[timeout-best-practices]]
=== Timeout best practices
This client is configured by default to operate like many HTTP client libraries do, by using a relatively short (30 second) timeout on all requests sent to {es}, raising a `TimeoutError` when that time period has elapsed without receiving a response. However, {es} will always eventually respond to any request, even if it takes several minutes. The {ref}/modules-network.html#_http_client_configuration[official {es} recommendation] is to disable response timeouts entirely by default.
Since changing this default would be a breaking change, we won't do that until the next major release. In the meantime, here is our recommendation for properly configuring your client:
* Ensure keep-alive is enabled; this is the default, so no settings need to be changed, unless you have set `agent` to `false` or provided an alternate `agent` that disables keep-alive
* If using the default `UndiciConnection`, disable request timeouts by setting `timeout` to `0`
* If using the legacy `HttpConnection`, set `timeout` to a very large number (e.g. `86400000`, or one day)