Detection of logging failure

0
For Dutch government clients, information security requirements are documented in the BIR - TNK. One of these requirements, 10.10.1, sub 5, states that when logging is no longer possible, the organisation needs to be alerted. In the context of the standard Mendix cloud, while using the logging module from the AppStore, this leads to a number of questions: Is there a possiblity that only the logging service fails, while the rest of the application keeps functioning? My current detection method is to write a test log message and retrieve that message from the database in a Health Check microflow. Is there a neater solution? What happens when my database grows beyond the assigned space (in theory, and attacker could keep spamming incorrect logins, which are all logged in the database)? Are there other scenario's which I should take into account?
asked
2 answers
0

Rom,

If the server is really busy (and about to start to fall over) it could be that the logging will stop while the application is still up (because the server is so busy trying to serve the users that there is no time left for the logs). But you will get notified about that situation because the application will be really slow and you get a warning notification that your application is not working normal.

Furthermore the logging is seperate from the database size. So your database size could be above the allowed space but the logs would still function (we had this situation once, but may be Mendix can elaborate a bit more about the finer details).

Regards,

Ronald

answered
0

Writing a test log message and checking whether it arrived via a Health Microflow seems to match your use case exactly? I wouldn't consider that a dirty solution at all.

Although writing logs to your application's database is generally considered a bad idea, you could log/check for logs on a separate lognode and filter that way? (ie: set the loglevel to NONE for all other lognodes)

answered