Where to Host a Fullstack Project on a Budget

welcome back I hope you had a good holiday if you celebrate Thanksgiving I have a pretty decent one had a weird sore throat problem but I'm starting to recover from it and feel a lot better now so that's good and I actually didn't have a big turkey I am half 11 knees so he actually had some middle-eastern food we had Kibby which is a minced meat dish mixed with bulgur which is actually quite good I enjoyed it a lot but today we're going to be talking about hosting your projects and where are some good options if you want to save money and deploy on a budget so the general way to save money when deploying things is to spend more time setting them up yourself rather than paying a service to set up stuff for you now the first thing that we're going to look at is cloud providers now they kind of break the rule because they do this thing where they have a juicy free tier so AWS has it where they offer a bunch of free stuff for their services for 12 months Azure has a similar thing where for 12 months they offer you free services and they also throw in a two hundred dollar credit for 30 days now AWS I believe also gives you like a hundred dollar credit you can get somewhere and all of these cloud providers have random promotions where they're throwing money at you also Google cloud has one where you just get straight $300 for 12 months and then they also have some always free stuff so these guys are kind of cheating because they don't actually charge you anything usually until after a year of using them or until you pass their free usage which usually doesn't happen right away so there's this illusion of value that you're able to not spend anything and get a ton of services and so this is the exception to the rule where you're able to leverage some of this stuff short term and that's the key thing these are great options if you want to set up a project that you think maybe take a year to set up or something you're gonna have a low traffic at first and then after that you see yourself growing the project and starting to get more users and hopefully generating some revenue through your service and then you can pay for some of the higher prices that the cloud services are going to charge because that is the way they function they try to hook you in with the free services you get set up and using them you're using their services so it's harder to migrate away and use a different service it's easier to just stay where you are so when the free tier is up you continue using them and you start using their paid versions of whatever you're getting from free before and this is where it starts to get more expensive than some of the other options but with that said I still think this is a super legit way to go especially if you see yourself in the short term having a smaller budget but in the long term having a more flexible or larger budget to work with especially because it's always gonna be cheaper to hire Google cloud to manage for example your database or your servers than it is to hire a human to monitor and make sure those things are working and so if you want to start cheap and then get some of the services that are gonna be handled for you I think a cloud free tier and then paying later is a great option to go with now there's some other things to consider when choosing a cloud provider - there's two things that I do not like about them the first is their complexity just to setup and use so they have gone to the point now where they are just supporting so many use cases and so many different services that their dashboards are getting very very complex and they are offering certificates to learn AWS or you get certified to show that you actually know how to use AWS now because it is that complex and you actually put on your resume that you are capable of using AWS because it is getting to that point where you really need to study it to fully understand how it works and make sure you're using it correctly and so that really ties in with the second thing you can really easily mess up and the price is not always super transparent it's very easy to spin up a lot of different services underneath the hood and then all of a sudden you have a larger bill than you were expecting so it's not always clear how much you're going to be spending when you actually start using these cloud providers so once you've decided you actually want to use a cloud provider the next question is which one should you use should you go with Google Microsoft desert or Amazon's AWS now this is a video and it's owned which one should you choose and this is also just not the only options Oracle has a cloud these just happen to be the biggest ones and I would probably recommend just starting with AWS because this happens to be the one that is used the most so it's gonna carry over this skill set to know a BWS is actually getting pretty important and a lot of different jobs need it and you'll use it in a lot of different places but I'm gonna leave an in-depth comparison for a separate video okay so that is option number one go with a cloud and take advantage of their generous free tears now option number two is one that I personally use and it's where we get more into setting stuff up ourselves and I'm gonna split this into four categories number one is hosting our database number two is hosting our server number three is hosting our front-end so server more are back inside and then our front-end side our website and then lastly what about user uploaded images now I'm choosing that because that's a very common thing that you usually have in your application and something you need to think about and affects how you host so starting with the database this is what I would not recommend using if you're on a budget and the reason for that is it gets very expensive very quickly and that is these hosted database options that are getting more popular so one example this is M lap where they will host a MongoDB database for you and you can see here it costs $15 per gigabyte and that is very expensive a gigabyte is not a lot of data but if you take it to the why this is so much it makes sense and that's because you're not just getting a database you are getting backups included for you you're getting support free transfer analytics hardware costs and maybe some other stuff thrown in there but it's something where if your database crashes you don't have to go in there and handle that stuff they have a team that is making sure your database is up and running and functioning correctly and so that can be really nice but it can be really expensive so another option and that's similar to this is with Redis labs where they are hosting a Redis instance for you and here you can see it is twenty two dollars a month for one gigabyte and for five of your bytes is one hundred and five dollars a month so very quickly with not a lot of gigs it gets very pricey but again they are doing extra stuff for you and again depending on your plan you get more stuff and so if you need these extra services maybe you want to consider these things but not everybody needs all of these things and it's one of those things where for a lot of sites if your database crashes it's okay to have an hour of downtime where you go in there and you fix it yourself that usually isn't too problematic for most sites it's just the fact that you don't really need 99 percent uptime all the time it's a very rare case that your database crashes at least it should be if it isn't you may want to have to consider switching your database or changing something up so what I like to do instead is host the database and a virtual private server so for example digitalocean is a company that sells these so is a vulture and there is a ton like this where basically you get a little computer and you can install whatever software you want in there and host projects and so it's also very flexible so not only can I stick MongoDB but I can also stick Redis on there and so I'm not limited to just hosting a single database on there but I can host multiple databases and I can also include my back-end server so personally what I will do for small projects is I will setup a single virtual machine which I will stick my database and I'll also stick my server on so they're close together now again this may not be optimal in all cases because if my database crashes so does my entire application but a lot of cases this is an acceptable amount of risk and I think is a great option if you are doing a small to medium-sized project and you see using this or building this over a long term because it is a cheaper price and if you want some of these other services along with just you know hosting your database you will need to set these up yourself which can be a pro or a con the pro is you can set up however you like and you get a ton of flexibility with it you get to choose the analytic tool that you use or you get to pick the way that you're backing up your database so for example backups were something that I really want it so what I do is I have a cron job which runs every day and it backs up the database and saves the database to s3 so on AWS on Amazon and so I have backups made every day and make sure I don't lose any data still and to give you a little insight into I actually like to use to manage everything on the virtual machine I like this open source software called dokku which works very similar to Heroku where you just push up your code and add in some database add-ons and you just let it handle the rest for you you know if they touch nginx or mapping subdomains and it's very easy to get let's encrypt set up so you have htps and so I actually have and you can run several different servers from the single virtual machine and so I have Postgres running in there and Redis and then also my back-end server and so being able to handle all those things in a single box or single server is pre cost efficient so how much do these virtual private readings actually cost well here is the pricing table for vulture so you can see it starts at 250 a month and it goes up from there and it really just depends on how much memory and how much CPU you need or you want now it's not just about those two things though and so that's one of the things that I naively would compare before so for example we can look at the price table here and the price table Oshin and we can see they're pretty comparable at least with the amount of memory and CPUs and transferred discs and that sort of thing that you get compared to each other but that doesn't mean you're gonna get similar performance and the underlying software that vulture and digitalocean use could be totally different and so to be honest with you I have no idea which VPS provider you should choose I'm personally trying out a couple different ones right now that try to figuring out which one I like the best I'm playing around with this one called SSD notes but I have no idea if they're any good but here is the thing that I think is the true test is to just do benchmarks and to see which one benchmarks the best is probably the best way to choose between these and so one place that does this there's this Josh tronic blog that compares them and so it gives you a nice side-by-side of these five providers and they'll actually go into CPU and some other things memory and compare the speeds of them and see which one you're actually getting you know you may pay $10 for each one of these which ones actually give you the most performance best memory or best CP or etc another place that I found that was pretty interesting was called VPS benchmarks where you could actually compare two different VPS and their speeds and whatnot so if you go to compare maybe providers where was I could actually like show suit compare VPS providers and so I could click vulture and I could click Google cloud compute or compute engine I mean we could actually compare these and get a breakdown of their both their pricing and also in each category they kind of rate them on performance and whatnot and you can even get into more fine-grain things down here on the performance at different levels so if you're into this thing and you want to dive into this you could send you could consider looking at these different benchmarks and maybe that'll help you pick but in this particular category I have no idea which one is the best vulture seems to consistently be one that across several different benchmarks performs pretty well and I've used it and it's been a positive experience so it's one I feel confident recommending starting with at least but there may be some better ones out there so let me know if you have a favorite on what you actually choose for this but in general price-wise they're pretty comparative and they offer about the same CPU about the same memory so it really comes down to the performance of each one of what they're giving you okay so the VPS will handle your back-end in your database now for your front-end this only really matters if you're not doing server-side rendering so if you're doing jam stack or you have a static website aka if you're using create react app using Gatsby something like that I actually like to host it on net liffe I if you are doing server-side rendering aka you're using Django Ruby on Rails using next yes something like that where you are actually you know rendering the pages from your server then you can actually just stick it with your VPS and you can do it there so this is for static sites I like to split them off and not render them in the VPS because you can get a really nice free tier from net liffe I and so it will take some load off of your server if you can serve stack files here so what I'll do is I believe it's up to a hundred gigabytes of bandwidth that you get for actually being able to host here and so that will last you a very long time and static websites are so easy to like migrate who you're hosting with it's very easy to if you reach over the hundred gigs which will give you give you for free it's very easy to switch if you want to go somewhere else but you may even want to stick with metal if I I personally just really like the experience with deploying here it's very easy it's very quick and I like it a lot so this is where I would go just totally free for front-end at your website if it's static so lastly let's talk about handling user uploaded images so I'm not talking about you know you have some assets on your landing page I'm talking about your Twitter your Facebook your some place where you want to handle the user uploading stuff and displaying their custom images and they could upload them anytime and so you want to be able to handle that so I want to first talk about the approach that I like the best and then we're talking about how you can't accomplish that approach so this is the one that I really like and that is to store wherever they upload store it at that size so if they upload an image that happens to be 900 by 900 you save it as 900 by 900 now the thing is your website may need it a different size or may need some different settings on the image so what you'll do is when you request the image you will give it some parameters for example the width and height that you need or maybe you need this width right and so what we'll do is the service will resize it get to the width that you want it on the fly then it will cache it and then it'll give you the image so this is something where the first time it may take slightly longer because it needs to dynamically create the image at the do size on the server and it's going to give you the image so the nice part of this approach is you have the original image at 900 by 900 and then you can resize it to whatever you want right so as your requirements for the website change or maybe you need different sizes in different places it's very easy to be able to just change the URL parameter you don't have to go generating a bunch of files ahead of time the right size that you want now I've tried doing this approach myself in putting this logic on digital ocean or vulture or one of those places but to be honest it is just a headache to try to do well it's not so much it's a headache to actually set up it's more of a headache to try to get performant and I was never able to get to a speed that was quite happy with I even tried seeking cloud fair in front but even then when image is loaded it was just a lot slower than some of the other options so I would actually recommend in this case at least I haven't had much success try and do it myself this is where I actually do like using a service to handle dynamically resizing images for me so there are two places that I know of that can do this image thing for you so the first one is cloud nari and they are great on a budget at least they're free tier handles quite a bit of images so you pay nothing upfront similar to how with the cloud provider has a nice free tier and then maybe you'll be paying more later but the hard thing I have with this service is or the hard thing to tell from this is they give you monthly credits so for example the free tier you get 25 and just a side note I really hate when these companies create their own little currency like this it's very hard to try to figure out how the currency maps to actual spending but anyway you get 25 credits which then used to a hundred or a thousand transformations one Giga storage or one gig of net viewing bandwidth you get the gist so what the credits could go to some amount of image processing so it's really hard to tell when exactly you would hit threshold to be paying $89 a month to be honest I'm not quite sure when a project you really need or how much usage you'd have before you reach the threshold because it's kind of hard to calculate ahead of time how much your image usage is going to be so that's my main beef with clouding area I think it's really nice I think the service is good but I personally haven't used them a ton because I feel like their pricing is a little less transparent than the second option which is image X so with this company how it works is you pay $3 per thousand images and then 8 cents per gigabyte of bandwidth and then you can transform the image with the property's height width and then they have a thousand other transformations you can do as me as many times as you want it doesn't cost extra so I have a very hard time trying to compare this pricing take Loudoun Aires and trying to figure out well with this many credits how much is it going to be on image X I'm not really sure so if someone's done the math do let me know but my gut instinct says once you pass the free tier of cloud and airy I feel like image X is going to be less money but to be asked I haven't no idea but I think both of them are pretty good options I've really enjoyed image X as a service I thought it's worked out really great but they don't have a free tier cloud Andheri does so maybe cloud nare just strictly wins and maybe it's even cheaper at the plus tier that was the case I probably should probably switch the cloud nari but I am personally using image X right now liking them but these are two really good options for handling your images but there you go in conclusion I like to use cloud services if I am doing a project where in the long term I see it making a lot of money but maybe in the short term I want to take advantage of the free tier otherwise I like putting my back-end my server on a virtual machine for example using vulture and then serving my website on net low Phi and hosting my images on either cloud Neri or on image X also one thing I totally forgot to mention Culinaire will actually host your images image X you don't doesn't actually host your images I forgot that's like a huge difference actually so I actually have my images stored on Google Cloud storage and then image X is kind of a cache layer in front of it so that's big thing to know difference between the two but yeah I've liked image X that's just a heads up so there you go that is how you can deploy on a budget you

Channel: Ben Awad

Views: 218,133

Rating: 4.9477262 out of 5

Keywords: Budget, Fullstack Project, Host a Fullstack Project, How to Host a Fullstack Project, host project, fullstack project budget

Id: Kx_1NYYJS7Q

Length: 21min 44sec (1304 seconds)

Published: Wed Dec 04 2019