Hi, sorry for the issue, the instance Redis.io runs on is very small and has only 1GB of memory, so the OOM killer killed the Redis instance. Normally this was not possible, because redis.io itself uses just a few keys, but lately I installed try.redis.io there, and apparently this uses a lot more keys, I'll have to monitor the keys usage and change the code in order to flush the old sessions faster. The current instance is costing 5$/mo at Digital Ocean, maybe I sent too cheap on this :-D But the idea was that I own everything and don't ask for expenses about Redis OSS official stuff, this was the idea back then. Now no longer makes sense in the new setups but still...
EDIT: just to be a bit safer I set a memory limit in the instance as well.
Btw if you see this as being not very professional, well it is :-D The website was setup by a friend of mine in an afternoon once we upgraded the site to use SSL. I have access and do sysadmin tasks without knowing very well how it is configured (it uses systemd for instance, while I setup things differently usually in my machines). The instance was configured without even a proper configuration file, just a few parameters on the command line. The thing is: we handle this as a "community" thing. And after all considering the near zero-efforts, and the 5$ instance for many many years, it kinda works well.
Anyway all the content is public and open source... (see redis-io and redis-doc repositories), and all our releases are tagged on Github. So the site is not vital in case it goes down for some short time.
EDIT: just to be a bit safer I set a memory limit in the instance as well.
Btw if you see this as being not very professional, well it is :-D The website was setup by a friend of mine in an afternoon once we upgraded the site to use SSL. I have access and do sysadmin tasks without knowing very well how it is configured (it uses systemd for instance, while I setup things differently usually in my machines). The instance was configured without even a proper configuration file, just a few parameters on the command line. The thing is: we handle this as a "community" thing. And after all considering the near zero-efforts, and the 5$ instance for many many years, it kinda works well.
Anyway all the content is public and open source... (see redis-io and redis-doc repositories), and all our releases are tagged on Github. So the site is not vital in case it goes down for some short time.