Coral UI: Request a new schema version button doesn't work

[michael-follmann]

What happened?

When I try to modify an existing schema for a given topic in the Coral UI (e.g. /coral/topic/delete-me/schema) and click on the button labeled "Request a new version", I'm not redirected to the expected page with the request form.

The link of the button looks good to me (e.g. /coral/topic/delete-me/request-schema?env=DEV_SCH&schemaType=AVRO) and the link works when I open it in a new browser tab. However, when I open it in the same tab, I can briefly see the expected page, but I'm getting routed back to the topic details (e.g. /coral/topic/delete-me/schema).

What did you expect to happen?

When I click on the "Request a new version" button, I'm getting redirected to the form for schema update requests.

What else do we need to know?

It only affects the new Coral UI. I can't see any errors in the web console or network console.

Requesting schema updates from the top ("Request a new" dropdown) still works as expected.

Also requesting the first version (e.g. for a new topic) still works. I only had this issue while updating schemas.

[programmiri]

Hi @michael-follmann, thanks for the detailed report! I'll see if I can recreate it 🕵‍♀

Can you tell me which version of Klaw you're running? Is it the latest one (v2.9.0)?

[michael-follmann]

@programmiri I can confirm that we're using v2.9.0. The following Docker image tags to be precise:

• https://hub.docker.com/layers/aivenoy/klaw-core/2.9.0/images/sha256-69a1f65aaf39e24bfd138a50a2592521d5bcda22309b22071fad4b4286cf27cc
• https://hub.docker.com/layers/aivenoy/klaw-cluster-api/2.9.0/images/sha256-347f71201c1943a49f030357494cdec119e3afd14c742a680c8b1b372a5fb397

[programmiri]

Hi @michael-follmann unfortunately I've not been able to reproduce your error yet. :/ Can you tell me what OS and browser you're working with?

It's an odd behaviour you're describing and since it works when opening in a new tab. From what you're describing and looking at the code, it sounds like it could be caused by some state in Frontend not being ready in the right moment (the state would be set anew when opening in a new tab, which would explain why it works there), but even with setting long timeout for api calls and throttling network, I'm not getting the error to happen. Which of course makes debugging hard 😅

I've checked the app with Chrome, Firefox and Safari on Mac OS, so maybe if you're using a different browser/OS, there's something that I'm not seeing on my system - I could check that in a VM in that case.

[michael-follmann]

Hello @programmiri, I was able to reproduce the issue on MacOS with a recent Chrome and Firefox version as well as on Window 11 with Edge.

Is there anything I can debug on my end? Maybe by increasing a log level or using a browser tool/plugin?

I created a screen recording with throttled network to make the loading behavior more pronounced:

I'm also able to reproduce the issue when I change the browser's URL to the link target (in the same tab). In this case I can briefly see the expected page before getting redirected back. I assume this is also happening in the first scenario, it's just too fast to see it.

[programmiri]

Hi @michael-follmann - sorry, I did not see your post here 😱 thanks for your patience! I'll get to that next week and see if I can reproduce it now with your infos!

In the meantime, if you haven't already, you could also check if that problem still occurs in the release we've done in May!

[michael-follmann]

Thank you @programmiri, the new release is highly appreciated! ❤️

I'm afraid I can still reproduce the issue in the new version (2.10.0).