Error importing a simple csv file on Data Visualizer

[josircg]

Kibana version: 7.10.0

Elasticsearch version: 7.10.0

Server OS version: Debian 4.19.152-1

Browser version: Versão 88.0.4324.182 64 bits

Browser OS version: Ubuntu 18.04

Original install method (e.g. download page, yum, from source, etc.): from apt

**Describe the bug:Trying to import a simple file on Data Visualizer
resultado3.zip

**

Steps to reproduce:

1. /app/ml/filedatavisualizer
2. Import attached file
3. I got "File structure cannot be determined"
4. I also tried to import fewer records with "Apply override settings > Number of lines to sample" but error remains.
5. If I import the same content within only 100 records on file, I got no error.

Expected behavior:

Screenshots (if relevant):

Errors in browser console (if relevant):

Provide logs and/or server output (if relevant):

Any additional context:

[elasticmachine]

Pinging @elastic/ml-ui (:ml)

[jgowdyelastic]

I was not able to reproduce this issue testing on 7.10.0, 7.12.1 and master
It could have been caused by a network request timeout, if the request was taking too long on a slow connection. This would explain why it worked with a cutdown version of the file.

[droberts195]

I got "File structure cannot be determined"

Was there any further information associated with this? Since 7.8 there should have been an explanation associated with failures, for example like in the second screenshot in #62316. Even if it's too low-level for you to understand, that explanation would help us a lot to narrow down what went wrong.

[josircg]

Thanks for the prompt answer folks.

It doesn't give any further message.

I don't think it's a network problem because it runs on a local network.
If you need any internal log or run a debug session, let me know. I have full access to this installation.

I tried also to click on "Apply override settings" to decrease the record numbers but it didn't work either.

[jgowdyelastic]

In the browser's dev tools, can you please find the analyse_file request in the network tab and let us know what the response looks like?

[josircg]

Fixed!

I got: Failed to load resource: the server responded with a status of 413 (Request Entity Too Large)

It was a NGINX problem! I had to increase client_max_body_size

I got also the message below. I don't know if this is useful to you.

Refused to execute inline script because it violates the following Content Security Policy directive: "script-src 'unsafe-eval' 'self'". Either the 'unsafe-inline' keyword, a hash ('sha256-P5polb1UreUSOe5V/Pv7tc+yeZuJXiOi/3fqhGsU7BE='), or a nonce ('nonce-...') is required to enable inline execution.

[jgowdyelastic]

ah ok, that's interesting.
Did you see any information in the analyse_file response? possibly something we could display to the user?

[josircg]

I think there isn't any response because server returns 413 http error.
If you got a 413, most certainly it's a nginx/apache limitation.

But I might be wrong. I don't know much how to debug on browser console or AJAX... I'm a server-side guy :)

If you need access to this installation to analyse the problem, I can provide you an external access and password.

[jgowdyelastic]

It might be worth me taking a look, I'd just want to see if there is a response body, because if there is we should probably show it to the user to help them debug the problem.

[josircg]

I will be glad to help. Contact me in private josir@irdx.com.br because I can't provide the username/password here.

[jgowdyelastic]

alternatively, here's a quick gif showing how to get to the network request for analyse_file in chrome.
I'd just like to see the response if there is one.

[josircg]

[jgowdyelastic]

Ok great, thank you for showing me the response. It isn't something we can use.
Thank you for your help!

[YevheniiPokhvalii]

I also had this error in Kubernetes. Added this annotation to the Kibana ingress:

nginx.ingress.kubernetes.io/proxy-body-size: "100M"