I am using triggers to send events directly to AWS eventbridge. Recently a bug has written NaN into a value and it has broken the trigger with
maximum attempts (30) reached processing event for trigger id=651217aaea3a6c3d84013d90: cannot marshal value fullDocument: cannot marshal value profile: cannot marshal value score: json: unsupported value: NaN
Not sure how to fix it…
I’m also seeing this come up and in the same boat as you.
Also for reference it turns out we had a bug setting an otherwise optional ObjectId field to ‘NaN’, but it still doesn’t reason why it causes this specific failure of Triggers completely.
We had a bug on our side writing the NaN which we did fix.
To get past the issue while it was fixed (and all events queued up), we had to use a ‘project’ on that trigger to exclude the field.
We weren’t using it in our handling of the response anyway so it was safe workaround.
It would be great to have another option though.
Chris