Error 413 - Payload too large error upon uploading files to weaviate hosted in Azure Kubernetes service

Support
1

Error 413 - Payload too large error upon uploading files to weaviate hosted in Azure Kubernetes service

2

weaviate version 1.22.3 hosted on azure lubernetes service AKS cluster

3

Hi @sg4544 ! Welcome to our community :hugs:

Is it in WCS or have you deployed yourself? in that case, have you used our helm chart?

Thanks!

4

I am trying to upload a file from front-end but if the file size is greater than 1 MB, I get this issue saying ‘413 Payload too large’.
The functionality involves user uploading any pdf file from the system to weaviate schema but the limit is 1 MB. Anyhting above this limit gives this error.

5

Hi, yes deployed using your helm chart.
helm upgrade --install “weaviate” weaviate/weaviate --namespace “weaviate” --values .\values.yaml

6

Please note only files >1MB fail the upload. So its ideally a config parameter somewhere limiting it.

7

@DudaNogueira Can you provide something regarding this?

8

Ok, just to make sure:

Are you uploading the pdf file to a blob field?

Because the way it usually goes is that you upload first to a backend, this backend will extract the content from that pdf, then chunk those contents and from the backend you insert the data to be vectorized and indexed.

Apart from a blob field, I cannot think of other scenario where a pdf/file should be uploaded directly to Weaviate.

By the way, have you seen the new Verba release? It has a really nice examples on how to achive this kind of feature:

9

@DudaNogueira
I am sharing the schema I have used:-

const knowledge_doc = {
class: ‘document’,
description: ‘a part of a document’,
vectorizer: ‘text2vec-openai’,
moduleConfig: {
‘text2vec-openai’: {
vectorizeClassName: false
}
},
properties: [
{
dataType: [‘text’],
description: ‘The text of the page’,
name: ‘content’,
moduleConfig: {
‘text2vec-openai’: {
skip: false
}
}
},
{
dataType: [‘text’],
description: ‘The document containing this text’,
name: ‘source’,
moduleConfig: {
‘text2vec-openai’: {
skip: true
}
}
},
{
dataType: [‘int’],
description: ‘This is the number of tokens for the text’,
name: ‘tokenCount’,
moduleConfig: {
‘text2vec-openai’: {
skip: true
}
}
}
]
};