All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
LET Providers used in Production
I'm so glad that I discovered this community and I've had the opportunity to try out a few providers; these are my favourite providers and the services I host with them:
Advin Servers
64GB (2022 BF Deal). Personal Blog, Main eCommerce Site, Social Network (Buddyboss)
32GB AMD Epyc Genoa (2023 Deal - Development Server) - Multisite (WP Ultimo)
GreenCloud and BuyVM - I Have two networks.
Network 1. Enhance Panel - A single control panel connected to servers in the EU, US and Asia; backup servers with BuyVM and their excellent block storage.
Network 2. Ploi Control panel with a load-balancing server connected to NGINX and Docker servers in the EU and US
I also use GreenCloud to host a BTCPay server, a CapRover server hosting many services and Hostbill.
Is anyone using LET providers in production?
Thank you to @Advin, @NDTN and @Francisco for their solid services
Comments
Crunchbits has been very solid for me, so is Racknerd.
Recently tried out HostEONS, pretty stable so far.
Ofc, Greencloud is top notch, I use them as well.
I wouldn’t use a single provider mentioned for prod
@Francisco being KIND of an exception, but it’s very limiting (especially to prod) not being able to scale, so what can go there is very limited for me (of prod stuff)
For me, prod = load may change
Personally, I’ve sold managed hosting for prod purposes for a while now. Since start, the most critical parts have always been at OVH
Backups have always been HostHatch & BackBlaze (double copies)
Then there’s the META stuff, that changes around ever so often. I don’t care if the control panel is down for 30 minutes for example, I’m the only one that’s using it for now
Edit: saw Crunchbitch was mentioned too, they seem reliable. Although that’s US only.
I wouldn't use LET providers for prod unless I'm running a fully stateless application or have my application database somewhere else.
I use my own Network for production but i also have Backups at Other Provider Network.
It doesnt Matter which Provider. LET, non-let or whatever.
The Message should be; never put all your eggs in Same Basket. And No, even a RAID10 can fail. Anything can happen
I believe most people wouldn’t
I think we always need to define how prod our prod is when discussing things like these
For me, if stuff go down, MULTIPLE people’s livelihood goes with it
Not HUGE money, but it’ll directly affect their salary and living
I’m not saying I’m a big fish, I suspect many people here are hosting much more important stuff
All I’m saying is that what’s acceptable for prod depends on what prod is to you
I appreciate what you're saying. The control panel that I use (Runcloud/Enhance) have a certain degree of disaster recovery, so I am comfortable using them. I cannot recall having any downtime with any of these providers and the price-performance ratio is pretty fantastic.
I honestly don’t think the difference in ongoing uptime will be particularely different
However, with one-man shows, IF a big issue occurs, they might be sleeping
Or if two nodes need attention at once, they might have to fix one first, and have the second node wait
Or if they have abusers, the ones they lease IP’s from might kick them and their subnet out (for those that lease)
Or if they generate abuse, their DC might cancel their rental servers or their colocation contract
Or they might just get a good job offer @ IBM and let it die
Just much more stuff that (in my opinion) can go very wrong
It's not really the same for me. I don't have a huge client base and not really critical to livelihoods. These are only the LET providers that I use. I use DigitalOcean for the higher end clients (although I suspect many in the community would have issues with DO!)
BTW @emgh - I appreciate your efforts at preventing the 'scalping' the other week. I only accepted the transfer because initially @HostSlick 'approved' on the post. The price that was being offered was really good, and I didn't realise I was being scalped until you 'disapproved', and then @HostSlick disapproved, and then it was too late, and I had to go through with the deal, not realising that @HostSlick was just days away from posting another banger of a deal!
Thank you anyway
@meditatingsurgeon haha no hard feelings, if I needed it I would have bought it too 😆
LET providers really do well when it comes to portable, high compute and high RAM workloads. If you need compute, RAM or even a (second or third) storage server for backups, LET pricing is great! The problem is mostly that the chance of deadpool is high and you really have no recourse for it.
"Nobody gets fired for buying IBM" is now "Nobody gets fired for using AWS". Sure you can find other providers with lower pricing but if the company disappears or goes bankrupt, you will be on the line for that decision.
The good news is that if your service is just a frontend to a database (like most web services are), just set up some docker containers and use Ansible to deploy. Put your database server on a DO/Vultr/Linode/AWS/GCP/Azure/Oracle Cloud cluster nearby and you have saved a few thousand dollars a month!
Modern horizontal scaling patterns have made it really easy to work with less reliable hardware (or in this case, providers) without much risk for downtime if something goes wrong. Use DNS load balancing between multiple providers and your customers won't even notice when something goes down.
@ehhthing thank you for your reply, really appreciated. EC2 and the Amazon relational database are something I have read up about and am really interested in. However, whilst setting up a Lightsail instance is a doddle, I'm having real problems setting up a test EC2 server with the free tier! I just can't SSH into it!. Not a VNC/security list problem as far as I know. When I manage to set up EC2, I will give the relational database a go - just for hobby projects mind you!
Docker is fantastic and I am learning Ansible too!
I do agree with it that Deadpool chances are high with LET Providers due to low margins.
Also even for mysql there is a cheaper option, just setup a mysql slave on another LET Provider so if your primary provider dies you still have your mysql slave on another cheap LET Host
The chances of two providers Deadpooling at same time is negligible
Of course always keep more backups even if it's any big provider
But this can save you even more ...
spartanhost
I haven't done much research into this but I think latency in MySQL replication is a issue for database performance, although I'd imagine if you were working with two providers that peer at the same IX in the same city it would be okay.
Very well said.
I think you've nailed it.
At risk of some self-flagellation, this is really where I see a strong use-case for LE* hosts (us included). A lot of the hosts here are not 1-man shops, and you are probably more likely to get an actual qualified human response without pulling your hair out or resorting to mutually assured destruction in the tickets. However: it's going to be slower if it's not an immediate service-impacting emergency. Some of the procedures might be less ironed out or fluid. You'll likely have less overall flexibility/customized front-ends. The flip side of that comes into play in a situation where maybe you need 1TB extra bandwidth temporarily because your project had some unexpected success. You at least have the ability to ask (and probably receive) with a LE-host.
Setup Master on same network and slave on other network so latency should not be an issue, if master fails then even your frontend fails, in that case the backup are there on slave, just make it master and setup another slave elsewhere
Agreed 100%
@OP please define LET provider. I mean, there is even @Hetzner_OL or @OVH_APAC on here but are they LET providers?
In any case i've never used AlphaRacks in production and i will not use it (or any similar host) anywhere in the future.
I use @InceptionHosting which I think is owned by @Clouvider for production from around 1 year and it has been pretty stable except for some small downtime issue once.
dude we're let users. we don't touch production environments
@totally_not_banned - true, no hard and fast definitions. I suppose the top tier are AWS, GCP, Azure, maybe Alibaba Cloud, and DigitalOcean. Not sure exactly where Cloudvider and Hetzner et al. fit in the 'league table'. I've never used (or heard of) AlphaRacks either. Why do you mention them in particular? Also, I have not seen Hetzner do a special for the LET community. I may be wrong, relatively new to the community. I don't know if that should be a factor in the definition.
@henix lol , but seriously I'd like to think that the aspirations of the providers here go beyond catering for the likes of hobbyists and hackers...
Sure, the term "LET provider" must be defined first. But considering all those small companies posting offers rather frequently here, their market share would look something like this:
So no surprise quite a few use them in production and I'm not among those. Yet I believe one might find some decent hand crafted niche services on LET. As for generalities - economy of scale rules there, so it's not possible for tiny firms compete with market whales, no matter if you put a comparison table of your services next to AWS, Linode on your site or not
AlphaRacks used to be the textbook definition of a bottom of the barrel shit provider. Nodes oversold to the point that their VPS/webspace was often times literally unusable, downtime, ignored support requests and generally zero fucks given in regards to anything. Their prices matched the unsustainable cheapness many LETers yearn for though, so they drew in quite a lot of clients (and produced a ton of drama because a lot of times nothing really worked but they didn't care or were unable to fix it because, well, nothing really worked) but still noone in their right mind would have touched them even if their products would have had negative pricing. I'm not entirely sure what happened to them in the end but they must have imploded somehow. From what i hear they have reincarnated though and are once again selling by the boatload. Maybe you can guess who i mean
Yeah, i'm not sure either if Hetzner ever did a real LET special. They used to be quite active though. Way beyond the mindless copy-pastering generic deals like they do nowadays. Guess they kind of lost interest in LET.
Here's what happened: https://lowendtalk.com/discussion/180381/some-juicy-court-drama-involving-alpharacks-racknerd-dustin/p1
@VPSSLIM netcup @berohost @crunchbits @HostSlick all are working for me without problems
Yeah, i saw that but it's not really about the demise of AlphaRacks. Did it simply deadpool overnight and disappear or what happened?
The only thing I'm running on LET providers that I would consider production is DNS and backup, because they are both very easy to replicate. I have 6 DNS servers hosted at 4 different providers, all in separate ASN's, and the backups are stored locally and on two different remote locations.
I do have like 30 lowend vps's but they are just used for testing, playing and idling.