Recently, our refresh_token started producing access_tokens that return "message": "Unauthorized request" for any graphql call made. I just tested creating a new access token through the CURL command in the docs to verify. These are inherently different than the token expired messages we are used to where we check for that case explicitly: if (message === "Bad token" || message === "Token is expired") {...,
We also have a developer account that we utilize for staging whose access tokens currently work. As a temporary solution, I have changed our production instance to point to this developer account.
I’ve also double-checked the scopes of both of the access tokens, and they are the same.
If this is any help, according to our error handling, we had this happen one-off on August 3rd, August 15th, and October 8th. Since those errors are coming from a specific cron job that is run every ten minutes, I can confidently say the issue resolved itself quickly each of those days. However, it has permanently been erroring for us since October 24th.
Appreciate the help. Did not want to post any sensitive account tokens, so just let me know what you need from me to further debug. Thanks!