If Amazon ECR is used the plugin returns only 100 tags (the first page). Because filtering is performed on the plugin side it may happen that no result are returned at all if the first page doesn't contain any of matching elements.

Without fetching and parsing all pages the plugin is unusable for cases when more than 100 tags are stored in docker repository.

There is related issue in github with PR just waiting to be merged : https://github.com/jenkinsci/image-tag-parameter-plugin/pull/48