In my custom GPT, I have configured my service to only allow requests originating from IP addresses listed in https://openai.com/chatgpt-actions.json. However, I’m still occasionally getting 403 Forbidden errors from IPs that aren’t listed in the json, but are very close to the listed ranges.
I’m wondering:
Is the IP list known to be unreliable or out-of-date at times?
Is the list updated automatically, or is it manually maintained?
What are the recommended practices for restricting IP addresses to enhance security in this case?