My ISP offers static IPs only on business plans that cost quite a lot more than the consumer rates. Difference is probably ~20€ per month. Plus I need the hardware and energy cost.
My VPS costs 10€ per month.
My ISP offers static IPs only on business plans that cost quite a lot more than the consumer rates. Difference is probably ~20€ per month. Plus I need the hardware and energy cost.
My VPS costs 10€ per month.
The “port forwading” thing is probably the biggest challenge. There is no “single tutorial” on how to do port forwarding. Some routers don’t even have a setting for it. And most people probably don’t have a public static IP, and I’ve seen provider fees that are so high that it’s cheaper to get a VPS instead.
And additionally, even if you manage the port forwarding, it is still some additional risk to host a public service on your own personal LAN.


Such bugs have already appeared in the past. It’s only logical that this happens from time to time, as the creators of bitwarden don’t test the compatobility with 3rd party servers.
What is important is that they’re not deliberately blocking vaultwarden.
The latest open weights model from google might be a good fit for you. The 26B model works pretty well on my machine, though the performance isn’t great (6 tokens per second, CPU only).


Does it offer notifications?
3 of your docker containers have new versions available
With portainer business, you could easily build an update procedure yourself. Just create webhooks for the stacks you want to update and run a daily curl script that triggers these hooks.
Similar approach here:
It doesn’t take more than 10 seconds to scan a doc this way.
They require an “data center” subscription now, and they will end support for that in 2029. So self hosting jira is basically not an option anymore.
No? Did I claim this?
The most upvoted post has ~20 upvotes, but there are a lot of posts with 1-5 upvotes.
I‘m not worried about myself. It’s easy enough to block stuff I don’t like. All I wanted to say is that counting daily posts without excluding bots doesn’t make much sense to me.
The instance is called lemmit.online, and the most upvoted post on the whole instance is “This bot is bad for lemmy”.
One thing that annoys me about each statistic about posts is that I don’t know how many of these posts are actually interesting and engaged with.
For example, there is a specific instance that just mirrors reddit content and has barely any engagement. The bot posts mulitple posts per hour, mostly without any comments or upvotes.
It seems rather irrelevant to compare these posts to actually interesting posts with a nice discussion and a couple of upvotes.
My suggestion would be to count and plot the number of posts that have at least a few interactions.


Is there a function to create a booklet or brochure?
This was a very useful feature to print a number of pages and have them in an easy format to read.
However, at least my Ubuntu print driver doesn’t have this feature, and I would need an extra tool to achive this goal.


I did it only once (yet) because i needed a specific addon for the software.
In my case, I wanted to use caddy webserver with a specific plugin. It was quite easy to create a new image exactly the way i wanted it.
I’m not a mbin user, but the combined frontpage looks interesting.


There are some nameserver providers that have an API.
When you register a domain, you can choose which nameserver you like. There are nameservers that work with certbot, choose one that does.


The only disadvantage I see is that all my personal subdomains (e.g. immich.name.com and jellyfin) are forever stored in a public location. I wouldn’t call it a privacy nightmare, yet it isn’t optimal.
There are two workarounds:


The best approach for securing our CA system is the “certificate transparency log”. All issued certificates must be stored in separate, public location. Browsers do not accept certificates that are not there.
This makes it impossible for malicious actors to silently create certificates. They would leave traces.
Yes, that is exactly what I meant.
I doubt a DNS filter would noticeably increase CPU usage. We’re not talking about deep packet inspection here.