Size limits for audio files

Thank you for your help, I have executed a request to IBM Watson in the Http Request node to process an audio file into text.
In that case, the connection with n8n is cut off when the audio file exceeds approximately 1 MB, as shown in the picture below.

What is the best way to deal with this?

スクリーンショット 2022-03-17 12.48.00

Hey @heiheihei, I am sorry to hear you’re encountering this behaviour. The error could suggest your workflow exceeds the available memory for your n8n instance (but you might want to verify this by taking a look at the respective logs on your server).

Seeing you are working with binary files here, could you try setting the N8N_DEFAULT_BINARY_DATA_MODE=filesystem environment variable? This would cause n8n to store binary files on the filesystem rather than keeping them in memory (make sure you are using an up-to-date version of n8n when testing this).

Thank you for your answer.
I am now using n8n in the cloud, where can I set the environment variables? From my screen, I have not been able to find it.

Oh, I am so sorry, I wasn’t aware you’re a cloud user. Cloud instances do indeed have very limited memory available, so I can see how this might have happened.

We should be able to enable this setting from our end, could you reach out to n8n.cloud support via the help center at n8n.cloud | Help center? There should be a contact form at the bottom. This will allow us to identify the right cloud instance for the change.

I see, I understand it very well. Thank you very much.

I will try the environment variable method as well.

Hi @heiheihei, the variable can only be set if you self-host n8n. For n8n.cloud we’d need to do this from our end, but I couldn’t find an n8n.cloud instance with your forum email.

So if you could reach out via the link shared in my last post or simply send my a direct message here on the forum we can sort this out for you :slight_smile:

I am sorry, but I would like to check one more thing before requesting that.

The following is the node you have created.

The rough process is.

  1. in DynamoDB, get the name of the file you want to download from s3
  2. from s3, download
    3.Http Request Node (speech to text), send a request to IBM watson with the downloaded file as an argument.

In this 3, the problematic situation occurs when the request is being sent out.

It could be considered as an error during the request, but could memory be the cause of the problem?

Best regards.

Hey @heiheihei, apologies I haven’t replied yet. The volume of requests these days is rather high, so it can take a few days before I get to take a closer look at more complex problems.

Memory limitations could definitely be the problem here (you are downloading a file from S3 which will be passed on through several nodes from the looks of it, each node will considerably increase the memory consumption of your workflow execution), but I won’t be able to say for sure until I know which n8n.cloud instance is yours.

Thanks for the reply.

This is the n8n.cloud instance I am using.

Would requesting a change to the environment variable still be the best way to go?

If you believe that changing environment variables is one of the best ways to do this, please make the change.
Is the only information required the own n8n.cloud instance?

Hey @heiheihei, could you share your n8n.cloud username or email address with me? Either via the support link posted earlier or via direct message to me? Otherwise I will unfortunately not be able to identify your instance.

I understand.

Thank you, again and again.

Thanks, I just replied to your ticket. You’d need to restart the instance for the change to take effect. The restart can be initiated via this button at the bottom of your n8n.cloud “Manage” page: