Hi,
In the past few weeks we noticed that our batch statistical requests are failing,
after long investigation we found that it is due to the <IDENTIFIER>.json that we add in the request_payload > output > s3 > url
Did anyone else encounter that?
was something changed in the API and maybe someone else know?
the support is not responding… 
Hi,
<IDENTIFIER> is a placeholder for feature identifiers obtained from geopackage file. These identifiers are now validated in the same way as non-templated part of S3 path key. Idenfifiers in your geopackage contain pipe character (“|”), which is not supported.
We’ll add a list of allowed characters in S3 paths to our documentation as soon as possible.