New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
Already had this in my personal notes to be added next. 😄 Now officially tracked here
They don't have an official Docker image, so I need to make one first. That's why it's not added yet.
Right. On the flip side, to me personally it doesn't matter much if I ssh into my VPS or if i log in to your web interface to click things. From my side I'd want it to be more convenient than self hosting, and the current pricing structure makes that a hard sell for me.
It's entirely possible that i'm just not the target group then as someone who can easily self host, but is just really lazy. Though, one would assume that the probability of any random person on LET being capable of self hosting approaches 1.
My target audience for this product are less technical users. There is a surprising number of people that won't run their own servers to listen to music or host pictures. I'm here because LET folks like free stuff and I need willing beta users. 😂
yeah, how long does it take to setup ssl for a custom domain?
A few minute usually. If you use Cloudflare proxying, I'd disable that first or make sure DNS validation works with it. Currently I see a few domains failing because of CF settings. Feel free to PM me the CNAME after CF proxying is disabled and I'll have a second look.
If I try to create a pod without filling in a name, it errors, but it doesn't error very visibly if you're on the resources tab. (I wanted to see if it would copy the application name or otherwise generate a value for me).
Can it be made to error like a missing environment variable, with the alert that appears regardless of which tab of the configurations you're currently in?
That's a good point. I thought the same while developing but didn't touch that code for a while. Will improve this point a bit. 👍
Hi, can you add these?
https://github.com/NginxProxyManager/nginx-proxy-manager
https://github.com/AzuraCast/AzuraCast
Those look fine. Added them here for now and will have a closer look later.
Currently I prefer apps with existing containers and low risk of abuse. Someone dropped a cryptominer into Jupyter last week. Not entirely unexpected.
I'm just a tad confused on this statement:
Which part exactly? Basically it's supposed to say that FOSS projects get X% of revenue while disregarding all the cost, if they do a collaboration. I didn't reach all the projects yet. Some are pretty hard to reach. Hence the invitation to get in touch.
In general it's pretty hard to give money to FOSS projects even if it's there: many people involved, maintainers can't issue business invoices, no reply by email, no reply via Github, maintainers don't want to fill tax forms (OpenCollective/Github Sponsors), fraud (Bountysource). To name a few I had so far. Fun. ☹️
Hmm, maybe that's where I'm confused.
I pictured this as something akin to Digital Ocean but focused on small ready to go server instances that you don't need to do everything by hand from scratch...
So I wasn't sure where that model comes in to play where the open source folks are getting paid, as I imagine their transactions and donations are handled outside of the platform
Maybe I just need to read it closer. I thought it was more of a micro servers to go for projects, but maybe there is a revenue thing I missed in there
Edit: OK now I think I get it. So we use the predefined open-source things you've picked (i.e. maybe it's nextcloud or something) and when you move to paid you will donate some revenue back to the open source communities... I really like that. I'd be in just for the social side of the project. My only challenge is I tend to modify the open-source projects and recompile so I can add features. So I rarely can use it as a vanilla package
Signed up will definitely play with it
The pricing is amazing, are you sure it will be sustainable? Also, are there any bandwidth caps/charges?
@m4nu Bug report: If you try to change to a subdomain that already exists, it deletes your previous subdomain & doesn't create a new one. If I try to change it to another one, it gives me a error that record "subdomain" is not found.
Good find! I relied on DB-level verification, but better to check before making DNS changes. Now updated to also give a nicer error message. Thanks!
Also started batching pod edits to make them faster and avoid locking issues. That should be a notable change when making edits too.
I'd definitely get an AzuraCast container, but that's because it's such a heavy system requirements that I have on a Hetzner Cloud instance that I'd like to drop.
Can't be heavier than PhotoPrism. They do ML-based facial recognition and eat more RAM than Chrome. 😬
AzuraCast will be added, since they already have good Dockerfiles. It just needs Redis, so I need to support that first. Currently I only do MariaDB and Postgres.
Playing around with it, and it seems pretty slick. Here's what i'm seeing:
With NextCloud there are some warnings in the Admin setup:
Also, if you're going to use resource based pricing, perhaps a "Shut down Pod" option next to the Delete/Restart buttons would be useful. Some containers might not need to be on 24/7.
Nextcloud is quite a beast. I'm currently using their official Apache container, which isn't great. Since it's a popular app I should probably look into improving a bit on that.
Even if a pod is stopped, it would still use storage. So most likely this won't influence billing. This is also how most providers do it when you stop a VPS.
Some of the more complex hourly providers like AWS/GCP/Scaleway have it worked out so that the compute resources (CPU, RAM) aren't billed while shut down but the storage still is.
I understand however there are internal reasons why this happens, like their reliance on block storage clusters to separate the disk from the VM and thus the ability to redeploy a started up VM to a different node, mitigating the issue of all the customers of a node running idle VMs and not making the provider any money off that node.
I don't know how your architecture is setup, though.
code-server would be a good addition.
I’m not sure how negotiations with these open source projects would go, and I’m also not sure if these are within your project’s focus as it’s oriented towards user apps.
Having said that, I’d like to see Hashicorp Vault and MySQL/Postgres as options, as I can spin up these quickly for experiments and maybe even as permanent data stores.
Great addition in fact! Already added: https://www.pikapods.com/apps
Billing working?
It's fully implemented but uses a Stripe Test account. So you can "pay" with any Stripe test card, like
4242424242424242
. Will use a live account in a few days probably. Working on adding a start/stop feature this afternoon.Can we lower the disk allotment below 10gb, for instance gotify container won't be using that much of storage
You can set it to 0 GB and not pay anything, if the app doesn't need any mounted storage.
This is now implemented. Hope it's what you expected. 😀
Good luck in your project, seems interesting ^^. Thanks for welcoming us during the beta!
Edit: borgbase is really awesome, i've understood you did it? Someone suggested to integrate borgbase as backup for our pods, i would like to big +1 that.