that does not matter you can just download more ram

restart: unless-stopped works for automatic restarting of crashed containers.

no wireguard should not decrease batterylife (see my other comment) we use wg eith always on without any problems. sometimes it stops on one phone but l9oking for the key icon and clicking the action button in the navigationcontrolmenuthingy is quite easy

that should not be the case because wireguard only ‘runs’ when it sends or receives packets. try setting the keepalive time a bit higher, 5 minutes maybe.

if you backup your vm data to the same provider as you run your vm on you don’t have an ‘off-site’-backup, which is one criteria of the 3-2-1 backup rule.

netdata is easy to set up and detects a lot of things on it’s own like databases and ntpd and…

pic from a google datacenter?

that’s also what we did in the early 2000s when building servers. today i don’t think it realy matters. i haven’t had a failed drive for about 10 years and only needed to swap them out because of the capacity…

i agree with you. also the ‘use vscode ssh-remote-eeit-plugin’ or ‘install a remote web ide thing to edit one file’ answers are… questionable.

syncthing does not require to open access to the outside and does what op wants.

deleted by creator

there is a famous quote i like to use for this kind of question: "in the end it doesn’t […] matter’. maintenance should be about the same (if you update manually, which is better in my opinion, because you don’t come home to something not working because of breaking changes in the software)

is the free vpn service the wireguard one? if yes and it is slow, than it won’t be any faster when using your own domain and exposing the server directly to the internet, because wireguard should be as fast as any direct connection. if it is not the wireguard vpn from the fritz box i’d recommend switching to it. this can be done by tge server jellyfin is running on if your box does not support wireguard.

in the end it doesn’t realy matter because you plan on running the services in containers. that beeing said, have you thought about arch?

it should not as it only ‘runs’ when pakets are sent through it, but it also should not be much more ‘expensive’ than normal network pakages. that is the best part about wireguard!

this is a bind mount, defining them like this in a compose file is for when multiple containers share volumes, you just need to write the name of the volume instead of the path, but…

 - database:/var/lib/mysql

is not that much less than

 - ./database:/var/lib/mysql

🤷

that is using bin mounts again. and i don’t get why this is discouraged, i actualy do like to know where my data is stored and don’t want to spin up a container that backs up using the volumes managed by docker. at that point you are using bind mounts again anyway.

regarding your original question, i don’t think it is possible (or even useful) to tell docker where to store the volumes for each compose stack because you are not supposed to access the from the filesystem directly anyway.

just continue using bind mounts

doesn’t systemd come with it’s own container thingy?

proper access points support 30+ clients without any problem (I doubt that the advertised number of 300 clients holds up for unifi aps, but 30 is definitely not a problem), especially for low traffic clients like iot devices.

why op gives up 2gbps from his 3gbps line is a mystery to me though…

of course /s

i guess the issue is not that you can’t use yaml but instead that a lot of things are removed from the actual config.yaml file. inuser tonhave all integrations set up in that file, after every update i get a notification, that this is deprecated for that integration and that i should remove it and use the gui setup instead. i have my config in git which saved me from needing backups, because all was already configured and could be redeployed at any time to another machine. now that is not possible anymore.