Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

MVPs and $100k AWS Bills: Reflections on our launch

octopus.com

Post reply on HN
Print view View original

91–100 of 101 posts

Re: MVPs and $100k AWS Bills: Reflections on our launch

#91

Post by cbm-vic-20 »

paulstovell wrote: post #59
LeonM wrote: post #28

Earlier quoted context omitted.

Even without the free trial, how on earth can you burn $100 in cloud resources in a month if you plan to resell it for $10?

The $10 was the starting price. It’s not uncommon for companies to have a cheaper “loss leader” tier.

A company should figure out what the "real" price should be, and account for the difference in the marketing budget. ie, if the service costs $100/mo to deliver, and if you would otherwise charge $120/mo for the service, but you're only charging $10/mo, you should consider that $110/mo a marketing expense if it needs to be a "loss leader".

It's got to show up in the books somewhere.

Re: MVPs and $100k AWS Bills: Reflections on our launch

#92

Post by redleader9345 »

plasma wrote: post #15

Octopus is a fantastic deployment product for .NET (happy user here).

I agree. I've used Octopus to automate deployments of internal LOB crud apps in the insurance space for a few years now.

My questions is, is it still aimed at .NET? I know it leans on Nuget as the package repo. But have they broadened to embrace deployment of other tech stacks?

Is Octopus considered best in class for what it does? Who are the primary competitors? Does Azure DevOps spell the end for Octopus?

Re: MVPs and $100k AWS Bills: Reflections on our launch

#93

Post by poxrud »

dclusin wrote: post #58
pjc50 wrote: post #19

> Octopus Cloud customers could start a free 30-day trial, which meant that those hundreds of trial signups per month, each of which cost us $100 to host, quickly added up. > We also didn’t have our pricing quite right. We initially launched Octopus Cloud at $10/month, with a different pricing model from the one we currently use. Unfortunately, this was one of the most painful lessons we learned because the deficit b…

It seems like napkin math could have alerted them to the potential worst case scenario costs. EC2 instance + db instance per customer alone made my jaw drop. Multiply this by 1,000, 10,000, 50,000 and costs are really regrettable. Edit: I know most database applications like Postgres can handle multiple databases. Does RDS let you have more than 1 database per instance?

EC2 instance + db instance per customer alone made my jaw drop

No, they mention in the article that they were using 30 databases per DB instance. Not sure why they used 30 when the RDS FAQ mentions[1] that for SQL Server you can have up to 100 databases per instance.

However if they would have used another database engine such as MySQL or PostgreSQL, they would not have had any limits to DB's/Schama's per instance and could have had an order of magnitude of cost savings. Using something like AWS Schema Conversion Tool[2] makes it easy to convert between different DB engines, and at a spend of $100K /month they could have easily afforded to pay a consultant to do the migration for them.

[1]https://aws.amazon.com/rds/faqs/ [2]https://docs.aws.amazon.com/SchemaConversionTool/latest/user...

Re: MVPs and $100k AWS Bills: Reflections on our launch

#94

Post by poxrud »

To ensure there was no way for one user’s data to mingle with another, each cloud instance had its own dedicated VM, database

No this will not protect you. Understanding the proper way to configure IAM roles with principle of least privilege will.

The author mentions that The issue here is Octopus is like a CI server - customers can run arbitrary code in the instance (not a problem most SaaS apps have). So a VM per customer was almost the bare minimum.

If you're using the same IAM role for your EC2 instances, and you're allowing your customers to run arbitrary code, well now at the very least they have access to everyone's databases.

Re: MVPs and $100k AWS Bills: Reflections on our launch

#95

Post by solidasparagus »

echelon wrote: post #13

I have a streaming multimedia app that I want to launch. It ingests from the client, does some transformations, then sends it back to them (or to other parties) in realtime. It's pretty novel though - Snapchat adjacent. I can't do it clientside. Nature of the problem. I'm writing it in Rust, so it's relatively fast. But throwing thousands of people at it will still require extreme scale. I'm going to use K8S horizont…

I would adjust my AWS service limits to prevent scaling past a certain number of instances. Set the limit to serve enough hypothetical users to demonstrate the demo is successful but prevents you from going bankrupt (better to load shed extra users than to end up with a bill you can't pay). If there is no number of instances that fits that criteria, I would consider rearchitecting to be cost conscious (I've heard positive things about Lambda-based services for keeping costs down).

Re: MVPs and $100k AWS Bills: Reflections on our launch

#96

Post by solidasparagus »

shiftpgdn wrote: post #74
tills13 wrote: post #66

Earlier quoted context omitted.

> For the cost of using EC2 you could hire an extra person to help grow your company. You _will_ be hiring an extra person to manage these machines. That's the beauty of The Cloud (tm). Need more compute? You got it. Provisioned too much? Get rid of it. If you're a startup, you don't have time nor money to be managing your own infra. Not to mention the ability to deploy to _any_ region within minutes. Now you're talk…

How much management does a bare metal box need? Nothing more than what an ec2 instance would need. There are colocation providers all over the world. Psychz.net is a great example. My assumption is a single box per region. There is no redundancy in your ec2 systems. You should google the Amazon profitability team if you don't believe me.

I can't find anything related to AWS and "profitability team" on Google. Can you share some info/links?

Re: MVPs and $100k AWS Bills: Reflections on our launch

#97

Post by syspec »

RantyDave wrote: post #77
LeonM wrote: post #28

Earlier quoted context omitted.

Even without the free trial, how on earth can you burn $100 in cloud resources in a month if you plan to resell it for $10?

The theory is that if you burn through a few (many) kilodollars doing it the inefficient way then you can at least prove the existence of a market for doing it properly ... and have a head start in actually having subscribers. I see the theory but $100k/mo in AWS fees is madness. They could probably knock a zero off just by deploying in containers.

Why is that? (Sincere question)

Re: MVPs and $100k AWS Bills: Reflections on our launch

#98

Post by solidasparagus »

mping wrote: post #51
shiftpgdn wrote: post #50

Earlier quoted context omitted.

I can purchase and then colocate 2x Dell R640s (8c/16t, 64gb of ram, as much NVMe storage as I can stomach) with tier 1 bandwidth billed at the 95th percentile and 24/7 remote hands support for about $150/month. The purchase price on the Dells is about $3500/ea. Total first year cost on this setup: $8,800 Total three year cost on this setup: $12,400 2x m5d.4xlarge EC2 instances (this isn't counting ingress/egress rat…

I'm pretty sure colocation is cheaper if you know what you're doing (what to buy, where to colocate, what kind of bandwidth required, etc). I'm also pretty sure a good portion of cloud customers don't know or don't want to risk it.

It probably depends what you measure. How do you scale up to a meet a relatively quick increase in demand? Will you lose customers during that time and how does that factor into your costs?

I think, at least for some business types, colocation will usually be more expensive because you have to overprovision or risk losing customers.

Re: MVPs and $100k AWS Bills: Reflections on our launch

#99

Post by solidasparagus »

johnnymangos wrote: post #56
DVassallo wrote: post #49

> We quickly learned that everything in AWS has some kind of service limit Except S3. Unlimited objects, and unlimited Put and Get requests.

Your comment is not true in all cases. S3 is limited to 5,000 get/put requests per prefix. So while it's possible to achieve what you've said, it's also possible to stumble upon this limit in a fashion similar to the original comment.

It looks like the number is slightly higher now - https://aws.amazon.com/about-aws/whats-new/2018/07/amazon-s3...

Do you know what exactly a 'prefix' means in this context? The first N characters? What is N?

Re: MVPs and $100k AWS Bills: Reflections on our launch

#100

Post by shiftpgdn »

solidasparagus wrote: post #96
shiftpgdn wrote: post #74

Earlier quoted context omitted.

How much management does a bare metal box need? Nothing more than what an ec2 instance would need. There are colocation providers all over the world. Psychz.net is a great example. My assumption is a single box per region. There is no redundancy in your ec2 systems. You should google the Amazon profitability team if you don't believe me.

I can't find anything related to AWS and "profitability team" on Google. Can you share some info/links?

https://news.ycombinator.com/item?id=19734261

Tangentially related: https://www.datacenterdynamics.com/news/ovh-ceo-unlike-amazo...

https://sellercentral.amazon.com/forums/t/amazon-literally-s...

https://fortune.com/2016/04/20/amazon-copies-merchants/

https://www.theverge.com/tldr/2019/9/19/20874818/amazon-allb...

Post reply on HN