Exactly this. As a low-level / embedded / non-cloud stuff dev, I've been getting up to speed through all the cloud-ification of the industry, but I'm still scared (not literally ofc) of running most things on my own on any big cloud provider (smaller ones seem more manageable).
I'm reading this and seems like being a customer of cloud services is like walking a dangeous path filled with gotchas and caveats, just jumping from cover to cover while hiding from danger, and hoping you're safe and didn't mess it up so far, "fingers crossed".
Like this tiny detail that he didn't realize was critical, so I would fall on it too plus on another 500s small papercuts: "oh I set cache up, so I hope all is well". "Yeah, no you aren't, I guess you didn't think of this detail about maximum cached file size! Gotcha, Game Over!"
Yeah cloud providers should have clearer communitacions and etc etc... but the fact of today is that they don't. So I'd never sleep well feelin 100% confident that I had covered and taken into account every minuscule detail and possible scenario that could end up being a disaster.
Another advantage is his big network that he can ask for help. There's also a chance that his blog post will reach the right person in Azure and he'll get a reduced bill.
As someone who doesn't have the same network or the "fame", I am concerned about what would have happened to me in that situation.
I am still waiting for a cloud without these dark patterns. But that will never happen because it‘s leaving a big amount of money on the table by not being hostile.
(edit) looks like that's not the case, I'm sure I used to have to buy a second instance a few years ago if I did want to use more bandwidth that was allocated
For those curious, the overage rate is $10/TB ($0.01/GB) after the transfer included in the plan.
The smallest amount of included transfer is 1TB for the $5/mo VPS.
If you use up your monthly network transfer pool, you can continue to use your Linodes normally. That being said, you will be charged $0.01 for each additional GB at the end of your billing cycle.
Funny enough...Oracle (OCI) makes it better, you can buy oracle"coins" 1to1 with $ and load your account just with what you think you need.
Hard requirement: My image can run on it (freebsd and linux), no proprietary BS, no special stuff, give me vm-"harware" make it fast, make it cheap, make it reliable, that's it..that's it.
And ATM i like oracle hetzner and vultr at most. If one of those change to my disgust i change, no big deal...just some dns rewrite.
This is how mobile and landline phone companies made enormous fortunes before flat rate billing. It’s called post-paid vs pre-paid billing.
"Predatory death-trap pricing" captures the spirit of the thing with rather more clarity. It is wholly intentional after all.
Or craft clothing for goths?
But the "dark" comes from its association with evil: "Defense against the Dark Arts", "The Dark Lord", "Turn to the Dark Side of the Force". It's a clear implication that the people are "selling their souls to the devil": knowingly doing something "a little bit evil" to achieve their aims.
I can see your point- if I'd never seen the term before I might have a similar reaction. But it's quite a common term now I think.
"Mr Politician what do you think about $big_co using dark patterns?"
"Mr Politician what do you think about $big_co engaging in predatory death-trap pricing schemes to defraud consumers?"
What is being asked in one of those questions is clear to everybody. The other is jargon and excludes the majority of the population from understanding the intended meaning. And note, you don't have to agree that it is "predatory death-trap pricing" at all. That is simply in the above sentence an accusation. Words have meanings. That the accusation be understood clearly by as many people as possible is important.
However regarding your other claims- I dont work on websites, I know the term "dark pattern". so your assertion that "nobody else knows what it is" is false. You might then argue that I'm still in tech, but that's just another goalpost that I can reach anyway: if you google for the term you'll find vox articles, explainer sites, even newyorktimes articles using this term.
So yes, it's a common place term and your assertion to the contrary does not hold any water.
As you say they make it hard deliberately.
Edit: Turn out Azure have this:
https://docs.microsoft.com/en-us/azure/cost-management-billi...
CMIIW, it'll be my first cloud provider if I can set one.
Not for actual cloud usage, like an actual pay-as-you-go plan where this would be useful.
https://azure.microsoft.com/en-us/support/legal/offer-detail...
It's very easy to overspend on these big cloud providers
Oh I can't edit the comment now oh well sorry if I've confused anyone.
My own instances go down all the time when I forget to ‘top up’ the account
Contract dispute cases might clarify it, but probably not in the direction any of us is hoping.
But the water company does not actually allow you to install proper taps to regulate the water so you use duct tape to do so, and due to an earthquake something falls on the tap causing your duct tape solution to fail leading to a massive surge of water, leading to your massive water bill.
Did the water company cause this? No. Your duct tape solution wasn’t resilient enough because it didn’t factor in an earthquake. But I would be justifiably mad that my water company does not allow me to install actual taps, and allows unforeseen and unpredictable situations to make me run up huge bills that could otherwise have been avoided with a proper tap.
This is very straight forward from their view, before: almost no traffic = almost no costs, now: huge traffic = $$$.
On the other hand, it doesn't seem that Troy did try to talk to them about this and seems to want to eat the costs himself. As it was his mistake. I think that's commendable. I also think with the amount of free advertisement Troy has done for them they'd be open to this and I can imagine we might see a followup post like "MS was so nice they waived my costs".
Maybe it's not a problem when you're dealing with millions of VC money, but there's no way in hell I would host anything in a bandwidth-metered cloud service when my or my own company's money is involved.
Eh, I don't know - either way he is a Microsoft Regional Director and MVP for Cloud (as well as security), runs courses on cloud deployment on Pluralsight, and has done speeches on Azure and reducing cloud bills, so if a he can get stung it doesn't say a whole lot good about my chances.
There are both spending limits and alerting that you could use, but would be impossible to predetermine from Azure's perspective, so they rightly ask you to.
The result is that you get a lowest common denominator type of dashboard. And hence a whole industry of providing just a prettier dashboard on top of AWS / GCP / Azure metrics.
Datadog started with a prettier dashboard for Cloudwatch data.
Cloudability started with a prettier dashboard for the Cost and Usage Report.
And also works the other way around. The individual product teams buy development environments to circumvent the console restrictions.
For example, a few years ago, the Redshift team purchased "DataRow".