Amazon Lightsail vs. DigitalOcean
blog.containership.io
blog.containership.io
Does anyone know if Lightsail uses the same CPU credit system as t2 EC2 instances?
Details on what CPU credits are can be found at: http://docs.aws.amazon.com/AWSEC2/latest/UserGuide/t2-instan...
Basically if you use too much CPU for too long, AWS will throttle your CPU usage. This is detrimental to performance in cases where you're not just idling most of the time.
I've had plenty of apps on DO use 40-60% CPU 24/7 without any performance degradation or throttling. Will Lightsail do the same?
It seems to confirm your suspicion. I'm not clear on whether Amazon implemented it this way on purpose, or if it was just the quickest path to launch a product, and they'll fix it later.
Edit:
I feel like they did this on purpose. It's tough to imagine an oversight as large as this slipped through by accident. It's also something developers will figure out in a matter of minutes while running on their platform.
Hopefully they only have it constrained now to make sure their system doesn't tumble over and they will lift it.
I guess they avoided to mention it because this pitch doesn't sound too attractive on stage: "Introducing Lightsail, a VPS competitor platform where you can get a 1 CPU core machine for $5/month, but understand that if you use more than a few percent of the machine's CPU for any extended period of time it will result in your CPU being automatically throttled to the point where your application will likely become unresponsive, and to add insult to injury we're not really making you aware of this during sign up. On second thought, just forget about it, continue using what you're using now because this isn't suited for production in most typical web applications that developers produce -- it's odd though because that's the market we're trying to cater towards.".
But at least Lightsail comes with a cape. https://www.digitalocean.com/community/questions/does-digita...
When I loaded the page the first thing I did was highlight that stupid box and kill it with the developer bar. I'm at 1080p and 125% DPI.
As a general rule, anything that can hover over other content on the page (from share buttons to menus and beyond) are going to annoy someone and break someone's experience. No exceptions. If it isn't happening today it will happen tomorrow.
Based on my experience with a portrait-oriented monitor for document editing, nobody bothers making websites work at viewport sizes other than "maximized on the designer's screen."
If you want me to put on my pedantic hat I'd that it's distracting still without the white space provided by 1080p. I really hate being overly picky like this but I want you to get valuable feedback. Ignore it, or take it but I applaud how quickly you fixed the issue to get this.
EDIT: I went back and checked closer and no there are no pixels that I can't see.
I loved Tutum until they became Docker Cloud and started charging per node. The flat monthly price on your product is way more attractive to me.
DigitalOceans main selling point is not only the price and speed at which you setup a node. The main selling point is their freaking awesome support and their tutorial/help pages that help a lot of people with troubles that might have.
AWS, with its amazing stack size, cannot compare in those 2 departments IMO!
Aren't these available for free?
Its an issue for DigitalOcean that if someones trying to install MongoDB on Ubuntu they can still use any VPS provider they want. The instructions that DO provides are not platform specific to them.
I still do want DigitalOcean to succeed though, I think there great and still love there product. I think building an S3 competitor(Not expandable volume storage) would make them more competitive. Its still going to be really tough though, right now any side project is going to be default on LightSail.
Over the years I've opened up a number of tickets on $5-20/month droplets for all sorts of problems. Even problems that were my own fault, which I received answers through either blog posts they linked me to or step by step instructions on how to fix it.
On multiple occasions they helped me recover from catastrophic errors like having SSH locked down to only accept connections with SSH keys but the SSH service died because I ran out of disk space. In this case they can mount a recovery disk to your droplet which will let you get at your file system.
We will stick with DO and Vultr for most things. They also still have a bit of a performance edge.
The danger is that we'll get to the point that >90% of Internet traffic in the US flows via AWS or uses something in AWS. At that point AWS owns the Internet and can start consolidating and shutting out other data centers and providers. It puts concepts like "peering" at risk.
That would be a monopoly. There would be only one way to host, at least for anything serious.
A hard oligopoly scenario with Google, Microsoft, and Amazon would not be much better. We want there to be many smaller players.
>TLDR: Amazon Lightsail 1GB will melt down with less than 15% sustained cpu usage
But AWS: If you get DDOSed you just pay a large bill. Which can of course get nulled if Amazon feels like it but still.
Data Transfer OUT From Amazon RDS To Internet
First 1 GB / month $0.000 per GB
Up to 10 TB / month $0.090 per GB
Next 40 TB / month $0.085 per GBWhich is a little odd as you can get Windows running on an EC2 instance (t1.micro) for $10.95/month. Although maybe the margins on Lightsail are too thin so Amazon doesn't want to pay Microsoft's licensing fees.
And before someone whines about Windows, keep in mind Windows Server 2016 now installs headless by default and will run on 512 MB of RAM (which is what the $5 Lightsail offering has).
Keeps the pricing down, since you're the one responsible for licenses.
The way Microsoft licenses Windows Server to Azure/AWS/etc is completely different to how they license it to you or I, and the pricing is hard to compete against at the lower ends.
I don't see it. Did I miss the performance testing section?
Performance comparison is coming in part 2. Sorry! Editing.
Pricing & Billing Gaurantees
should be
Pricing & Billing Guarantees
I'm not sure how reliable they are but if they are this doesn't seem good for AWS:
>TLDR: Amazon Lightsail 1GB will melt down with less than 15% sustained cpu usage
>
>Okay then, experiment aborted. 25% sustained cpu usage is too much, all other VPS at the same price can handle that but maybe we're just over the limit, we moved traffic to another machine and let the Lightsail VM cool down for a few hours.
>It should be noted that for $7, Vpsdime easily handles that traffic.
>For less than $4, OVH "VPS SSD 1" can do the same.If they leave it as-is, I can imagine anyone using it for something non-trivial being highly disappointed.
AWS has been saying they will open in Canada for quite a while... seem to keep pushing the date further and further.
This looks like it will be it. They just got so annoying it was unbearable.
One small correction, AWS doesn't have any more the largest footprint when it comes to regions and revenue. Microsoft Azure now has more regions and DCs than AWS and Google combined. And for the revenue they reached 15B dollar mark by last quarter, making it the the largest and the most profitable cloud provider ever. Who would thought that is possible two years ago!
While AWS is a great collection of services, it is needlesly complicated to use and at times very frustrating. Using it at work is fine since the span of services is so wide I can forgive the awful user experience, but if I have to vote with my feelings, I choose the service I use in private, when I pick what I like.
In this case DigitalOcean.
digitalocean is being attacked on all sides, interesting. Hard to believe that giant AWS is trying to imitate one of its upstart competitors...