My users have been getting this error UT005028 followed by a Proxy unreachable/crashed, stopping it now, it happens right after the proxy is activated. It is somewhat intermittent, eg some users get it others don’t when it happens, they can’t access the application. It seems that it can be remedied by deleting cookies or using InPrivate/incognito mode.
UT005028: Proxy request to /proxy_endpoint/######/dt-ext-fixedheader-1.10.20/css/fixedHeader.dataTables.min.css failed
I can’t tell if it is related to an application, networking or security issue. I know datatables have compatibility issues. It might just be an issue with http requests though because datatables are the only ones that use http and ajax.
Are there any ideas how to trouble shoot or log this?
My users also experienced something a while ago but I ignored it initially as it was very rare and not reproducable.
Btw the error I get is the following:
2024-08-02 13:26:47.805 ERROR 1 — [ XNIO-1 I/O-5] io.undertow.proxy : UT005028: Proxy request to /proxy_endpoint/206ce73e-d2da-4f7e-8637-b6a7ba6cd5ce/grid_data failed
Now recently it happened multiple times in a row.
Is there any solution to this already?
Ok thanks, I’ll upgrade in that case.
@tdekoninck, I tried using version 3.1.1 but I came from a snapshot version of 3.1.0 (3.1.0-SNAPSHOT-20231002.085346)
I now get the following error:
Included health contributor ‘readinessProbe’ in group ‘readiness’ does not exist
Has this been deprecated or given a new name?
This is were its being defined (as part of the redis-sentinel setup)
management:
endpoint:
health:
group:
readiness:
include: readinessProbe,appRecoveryReadyIndicator
Hi, I suggest to remove these lines completely, the default of ShinyProxy should work fine.
Now it seems to not be able to start the redisContainer bean, I will leave out the complete stack trace but it essentially comes down to the redisContainer not being able to start
2024-08-05T15:10:32.792Z WARN 1 — [ main] ConfigServletWebServerApplicationContext : Exception encountered during context initialization - cancelling refresh attempt: org.springframework.context.ApplicationContextException: Failed to start bean ‘redisContainer’
So far, the issue has been resolved with ShinyProxy 3.1.1 which has been stable and in test for a couple of weeks as well as moved into production for a week.
This sounds like the more general closely related overloaded cookie issue. We have to carefully remove any excess cookies in our router because certain R packages don’t handle large headers well (datatables argh). Not really related to ShinyProxy.
1 Like
I still get complaints from users that are experiencing crashes even after the upgrade to shinyproxy 3.1.1.
I still see logs with UT005028 error codes popping up followed by a Java stack trace.