yes, we know the implementation isn't prefect, we have some issues with unselect, since it works peaty much the same in Gmail, we felt it was good enough

as long as the 3 services are subdomains of .my.domain this should work

192.168.1.83 is a local network address, is it accessible from your browser?

API server isn't aware of fileserver location, it's clearml-data responsibility to write the correct path in the server.

Hi Manish, did you check our FAQ?

AgitatedDove14
in the browser a NaN will crash JSON.parse(), so we don't have them on graphs.

Hi @<1888759776226054144:profile|MistakenFox18>
can you upload the response of the last tasks.get_all_ex query?
you can find it in the browser devtools (F12 opens them) on the network tab.
also what is your monitor resolution and scale factor?

it might be an issue with the way clearml.conf is setup

Hi CloudyHamster42 , we have identified the cause for this issue and currently working on a fix for the next UI release

can you please also report in github so we have an item to follow once the fix gets released?
https://github.com/allegroai/clearml-web/issues

I'm looking for the url of the image coming from the api server

MotionlessCoral18 if you build the UI, instead of overriding stuff with hard coded values, you can set fileBaseUrl in src/environments/environment.prod.ts

probably need to set
hovermode = 'x'in layout

SillySealion58
something look odd with the bucket settings.
url has port 443 which usually means ssl secure, but the secure flag is false

Hi Ryan, just wanted to update, we found a performance issue with a 3rd party library we use for a detailed compare of text. we added a limit on advanced compared, so the UI won't get stuck. the fix should be included as part of the next server release.

I'll need to check how to configure the UI for this type of deployment as the UI has just a few configurations it can guess the files server URL from.

Hi ResponsiveRabbit81 , are both resources on the same protocol security level http/https?

i'd focus on nginx or whatever azure infra you have between nginx and the client.

can you try running the curl request from within the web container?

one thing I can think of is nginx default client_max_body_size
limit
None
800KB is about 1MiB
can you try altering nginx configuration in the running web instance and reload the new config?
nginx -s reload

you can manually override this configuration by adding or updating (if you have 1) the configuration.json file next to the index.html
with fileBaseUrl

it might have some mixed or insecure content blocked

yes thanks, that would be very helpful

UnevenDolphin73 do you see a "Session expired" message on the top right in these scenario?

Hi JitteryCoyote63 , shouldn't be too hard, but we would probably need to check what most of the users prefer.

Thank you Ophir, managed to reproduce on our environment. we'll work on a fix.

Hi @<1523701366600503296:profile|GentleSwallow91>
https None is not a valid URL.
you're missing a colon : after https

on second thought if direct request to api server works while reverse proxy fails, this points us only to nginx.

WearyBaldeagle85 you need to navigate to the plots tab