Trello API Internal Server 500 Error

Question

I have a data pipeline where I pull data from Trello's API every day. I keep getting an internal server error 500 when pulling cards from certain boards. This points towards it being an issue with Trello's API itself. Anyone else experiencing this or know how I can mitigate the risk of it?

I have been experiencing this for a couple of weeks now, before the larger error that happened last week.

Thanks

Jacob Barr · Answer

Hi Arkadiusz,

Thanks for your fast response :)

For context, this pipeline has been running successfully for the last 6/7 months with no issues. The boards that it has failed on are also some legacy boards which haven't had data added to them in months, so it can't be that there's too much data on them, as the amount of data on them hasn't changed in months but it's only been an issue the last few weeks.

Also, I'm not pulling actions in this part of the pipeline that's failing. I'm just pulling all the card details on the board using the url:

https://api.trello.com/1/boards/{board_id}/cardsI will try to narrow down the fields that I export and see if that helps, as currently I'm pulling all fields. Also, sometimes it fails and sometimes it doesn't! So I think considering this context it more so points towards it being a Trello-side problem

Thanks

Arkadiusz Wroblewski · Answer

Hello and welcome to the Community @Jacob Barr

A 500 error usually means Trello is breaking on the backend, but since this is only happening on specific boards, it’s highly likely those boards are just too heavy for your current API call. Trello's database can easily time out if you're trying to pull a massive payload like cards packed with actions, attachments, and custom fields all in one go.

Try stripping your query down to the absolute bare minimum, just pulling basic card IDs and names. If that actually goes through, you'll know the board is just too large for a single massive dump. You'll want to adjust your pipeline to fetch heavy extras like comments or actions in separate, smaller batches.

Test that and we can go from there.

Best

Arkadiusz☀️