Good morning
Through a discussion in another forum that I use, I came across the topic of data protection in Discourse installations. The software that has by now become the de-facto standard for forums is actually quite data-hungry with its reading lists etc. Users in that particular forum had checked what data was being stored about them, compared it with the privacy notices, and were outraged. Something then happened and the forum was offline for a short time. Since I can no longer find the discussion, I suspect there was a complaint filed with the relevant supervisory authority and the forum admin then acted urgently in response.
Since I generally advocate for data minimization, I was interested in how other forums handle the topic of IP addresses in particular. By default, Discourse stores the IP of the registration and the last X logins indefinitely, and even with the „anonymize“ button for admins, this data is apparently not cleared — at the latest that’s where you’d end up with an (unintentional) data protection violation. I can’t judge this myself, since I don’t run a Discourse instance myself where I could check this (as a user, you logically no longer have access to an export of your own).
In an export of my data, I unfortunately found exactly such IPs, of which at least the registration IP is older. In my case, this isn’t a problem right now — but I can’t understand why the IPs need to be captured at all? Especially the Digital Services Act actually wants to enable anonymous usage, if there are no substantive reasons against it.
There are apparently scripts for Discourse that regularly truncate the captured IPs (e.g., to a /16 for IPv4 or /48 for IPv6). From my point of view, such an approach — if one doesn’t want to set the X-Real-IP to something fictional via an upstream proxy before Discourse even sees it — is a balanced solution.
Maybe the tech team here can provide a remedy and, for example, provide such scripts that clean up the captured IP addresses.
In my own export, I found unnecessary IP addresses in the following fields:
account.ip_adress
account.registration_ip_address
loginHistory.client_ip
The following fields were (in my case) empty / possibly blocked:
topicViews.ip_address
topicLinkClicks.ip_address
profileViews.ip_address
Due to lack of data, I cannot assess the following field:
history.ip_address
I’m posting this here because I suspect that this data collection isn’t actually intended to be permanent in the long run and is simply due to the extensive default settings. If this data is supposed to be used to combat bot registrations or similar issues, then at the very least the privacy notices should be amended accordingly (in particular stating that this data is retained permanently and whether this data would be deleted upon deletion requests). I’m now assuming that sharing this data with third parties is already prohibited.
Thank you very much
Dezentral