Perhaps implement some type of “password canary” - some type of test account(s) with known high-entropy passwords.
Have an automated system send periodic login requests (or any other requests which contain sensitive information that shouldn’t be logged) for this account, and have another system which searches log files for the password.
If it’s ever found, you know something is leaking.
Have an automated system send periodic login requests (or any other requests which contain sensitive information that shouldn’t be logged) for this account, and have another system which searches log files for the password.
If it’s ever found, you know something is leaking.