Hi @GrandMa, welcome to the community and sorry for the trouble.
Based on your description it very much sounds like you’ve hit an out of memory situation. It might be worth checking your server logs to confirm this.
Seeing your are fetching hundreds of file here you can move some of your workflow logic (such as the file download and upload parts) into a sub-workflow which would then only return a minimal item (such as the next page URL)? This would mean the memory required for each website downloaded becomes available again on the next loop.
I’ve posted an example for such sub-workflow over here. It’s not exactly a copy of your use case (but rather paginates through calendar months), but it hopefully still helps with getting started.