Open Andrei-Predoiu opened 1 year ago
Thanks for reporting, would you mind also opening a support ticket to track this please?
Thanks, i opened a case
Hi Tom, microsoft informed me that they raised the case as a bug.
@tomkerkhove any updates or eta?
No, sorry.
What kind of load balancer is fronting this self-hostd gateway? I see it runs on Google so I am not sure, but on Azure the maximum timeout is 240 seconds and we're usually seeing the request to end after 230s because of this
That's a great question, we have nginx with 900 seconds timeout but there are several network components after nginx, until the traffic reaches the "internet". I'll look into it.
Can you please try applying the following settings and see if it resolved it?
Report
On version 2.0.4 the Self hosted GW there was an internal max timeout of 240 seconds. Newer versions have an internal timeout of 230.
This is not documented and even seems to contradict existing documentation relating to the
<forward-request timeout="..."/>
policy.Expected Behavior
The GW timeout respects the timeout set in the
<forward-request timeout="..."/>
policyActual Behavior
The GW times out after 230 seconds
Steps to Reproduce the Problem
http://httpstat.us/
http://httpstat.us/200?sleep=500000
Logs from self-hosted gateway
V2.0.4
V2.3.1
Self-hosted Gateway Version
2.3.1
Deployment
Self-Managed (YAML/Helm)
Platform
Google Cloud
Kubernetes Version
1.25
Anything else?
No response