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.
[Show LET] Idlers porn
Hello everyone.
Long time lurker, first time poster.
Vibe-coded recently a simple monitoring dashboard for my fleet of not-very-idle idlers, because Grafana was to complicated to learn.
Most of the nodes are $7/year class, some are highendtalk like Vultr. None of them are sponsored, there are no affiliate links, just some info in addition to yabs people post there, since yabs is static and cpu steal vary.
I plan to publish latencies and packet loss in a few weeks. Feel free to roast or suggest improvements: https://lowendnet.com/
Comments
Nice one
GL with your project. I like that you have included provider and location.
EDIT: btw is that real load i see there ? Or those vms are lxc ? I doubt netcup is lxc..
@letslet you going to open source this project?
@bdspice must be professional porn idler
some idlers below
Your Axus went offline
EDIT: came back
Both servers are having the same 'issue', it is a 1 minute timeout, might be djunior firewall blocking those packets
I'm about to say something heretical but... what's the point of 110+ low-spec idlers? I can understand running a private VPN cluster - I do that too. But most servers in the "$7/year class" as mentioned in OP are not going to be in exotic locations worthy of a VPN. Nobody really needs 30 VM exit nodes in Chicago.
I suppose if you've got clients who you're hosting websites for, then that's a different story, but otherwise, at $10/year on average you're paying $1100/year total. For the same price you could get a current-gen server like a 9950x which will provide you more capable compute which will be much easier to manage, or you could take a serious step down and avail yourself of ~5 of those $200/year dedi deals to get geographic diversity.
I did this calculation last year and dropped a lot of my VMs in favour of consolidating to more capable dedicated machines for the same cost.
The only reason I see is to get diverse IP addresses for cheap, or to learn how to manage a fleet of servers for a CV/porfolio. Or perhaps to use for multi-perspective monitoring of some kind.
In any event, congrats on the project. Can't believe vibe-coding something was easier than learning grafana, but congrats on the project and would love to hear what the actual usecase is!
Is that impressive list of VPSes entirely real, i.e. that's the machines you actually own?
100%. over 5tb porn is idling which i never saw.
ideal customer
Missing a "Go to Top" button
Probably the champion in the "CPU Steal" category is Contabo, but they don't even have a single VPS in the test right now. And Dasabo is currently the one with the highest CPU Steal,
Love your tool good work! Let me know if you think about releasing it!
I’m seeing one of them briefly go offline in his dashboard, which is odd, it comes back online almost immediately.
@tfgp99, if you’d like us to investigate, please open a ticket. 🙂
I used to think like that and now I have a hefty dedi bill as well as lots of idlers!
cool project, i like the cpu steal being displayed
mr.finance is speaking. Having 100's of idlers is so much more fun!!
Impressive! I thought I was nerd, but this is uber nerd.
Respect.
idlers unite !
Can you tell us more about how it works? I'm assuming a script that polls into a central location?
I prefer the opposite, pushing my VPSes hard.
Thank you, i've opened it (#ADT-353552)!
Thanks! The load is real and low recently. Those are KVMs with Ubuntu 24.04 or Debian 13.
No, not really - too tied to the existing infrastructure. There are other great projects, I just wanted slightly more custom thing.
Just a suggestion, when you hold your mouse on a CPU Core, could show what CPU is.
Actually this dashboard is pretty cool and clean!
You are totaly right, there is almost no point for normal people. Except I'm getting geographic, IP, AS diversity, which is beneficial for accessing rate-limited resources (and cheaper, than rented proxies, I believe), having latency measurements independent from RIPE Atlas for hobby research project and also having ~20 BGP uplinks for the same reason, as well as running honeypot servers from time to time to see trends in various scanners.
I do have beefier hardware that handles actual workloads (not on the dashboard), but having at least some compute "at the edge" allows reduce traffic on these servers - the DNS scanning task, instead of firing each request via proxy, is originated from the VPS and batch results are getting compressed and only then uploaded.
The learning part is also important - terraform manages ansible's inventory, which configures and deploys consul with nomad. Configuration management and service discovery via Consul, task orchestration via Nomad, distributed work queues, not the fanciest stuff, but enough to get the feeling and to learn how it works in general with muliple problems along the road.
I'd say rent, not own.
My "usual suspects" in overprovisioning cpu steal are DasAbo, DeluxHost, MassiveGrid, SoftShellWeb, but I don't feel entitled to complain, considering the price.
Terraform manages ansible inventory. Ansible configures OS, hardens, adds node to wireguard mesh, installs docker and sets up consul and nomad. 3 consul clusters (americas, europe, apac), one nomad cluster. Nomad runs, in addition to everything, prometheus node exporter and prometheus server discovers all the nodes via consul and collects metrics. This web page is basically a frontend to prometheus + nodes discovery via consul and mapping to human-readable locations and ISPs.
That sounds suitably interesting! I haven't touched on Terraform yet for my infrastructure, just getting into Ansible for now.
I'm in the same boat with some of my dedicated boxes. Cheaper to rent a new server than to pay for extra bandwidth in some locations.
I reckon you might be right so far as proxy pricing is concerned. We have two providers who frequently run proxy deals here on LET, their pricing tends to be $6-$12/month for usually 3 different subnets, but that'll al be under the same ASN.