I really like their take on, essentially, “don’t drop everything just because there’s yet another CVE”. There being vulnerabilities is the “normal” under our current development of technology. The core is making it a non-tiring normal. When I hear there’s a new CVE I give a quick read to the minutes. If it can’t pass the test of working beyond “requires a user already logged running a specific program from /tmp instead of $PATH, with the computer facing left and the brightness of the monitor set to 72.3% o a Tuesday”, then I leave it for the nightly cronjob to upgrade the container, and even if that somehow does not work then I would be checking tomorrow anyway. The higher-ups are not assigning a fair paid overtime for this, or even any at all, so.
Thing is, like with the recent php, you don’t have the info you claim to base your decision on. Currently it’s ‘any nginx with http2’. Bet you, next week we know about the ‘computer facing left’ detail. But today we have to assume the worst.
We have been building layers of security for years. CDN’s can quickly block lots of things https://www.ioriver.io/blog/cdn-security-risk
If I was a security researcher and found an issue my first call would be the developer, my second would be some top CDN’s so they can search for and mitigate traffic in web applications.
Edit. I should also add docker and fast updating. I just updated everything I Run with internet exposed ports in about 20 minutes.
Maybe updating everything brought in a new supply chain attack and you just infected yourself
I only run updates that are old for this very reason (unless there’s a very specific cve fix)
It really is an issue for user driven repositories like aur and all those pypy, gems, composer stuff. For distribution controlled packages the risk is way lower.
Or maybe you just got hacked because you didn’t update.
This line of thinking can go back and forth forever.
See above where I said if it’s for a cve it gets installed asap



