Do all the things like ++ or -- rants, post your own rants, comment on others' rants and build your customized dev avatarSign Up
Low end is someone who doesnt value their skills and you shouldn't either.
Upper end seems the high end of correct. They seem to have the timeline right and estimated 2 months at $200/hr.
I would scope a contract with deliverables and a time frame. I would say all of that should take less than 2 months if it's just what you describe.
Cost, 100 to 200 freedom bucks an hour sounds about right for a short gig and a top level resource. Low end is going direct, high end is through a firm assigning a resource.
Generally the longer the engagement the better the rate. Shorter engagements end up paying a fee assuming a dry spell at the end of it.
dontknowshit18724dI’m not an expert on AWS but I do have experience with all the products you mentioned.
The complexity of what you require, really depends on the state of what you already have and how much work needs to be done to integrate.
I would say in an ideal situation it should take no more than 2 weeks of work but like I said it really depends on the state of what you already have, it will also require on going monitoring and maintenance.
I would suggest that if you have the ability to build a product that requires this, you should just learn how to deploy it, it not difficult.
The question I would ask myself is, Do I even need AWS at this point?
sariel26123dTerraform will be your friend
zymk111723dI hope this info might help,
I'm an AWS Certified Solutions Architect (Associate) and I personally charge about 80-125 an hour depending on how big the project is. If there's less than 160 hours of billable time on a project I would try to aim for 100-125, and if it's over 160 billable hours then I drop the price down to 80.
I work in AWS everyday and depending on how production ready your code is will determine how quickly you can setup infrastructure in AWS to serve it. Without knowing the finer details of your application code, or database requirements, or networking; it's a little difficult to say exactly how long it would take. Scripts shouldn't be too bad either. About 3-4 weeks would be a fair estimate. Make sure they understand AWS costing.
I'm working on a large project now to reduce my company's AWS bill. I've managed to save a little over 9k a month. You will want an engineer that can make sure your monthly bill is as low as possible. AWS gets really expensive, fast!
chowdercake142223dI have some knowledege with aws and most of you the things you mentioned may not even be required.
Like, you don't need zerossl nowadays cause let's encrypt provides it's own certbot for linux machines.
If you're going for auto scaling I'd recommend using spot instances (a lot cheaper 70% of on demand).
Cloudflare(free version) never gave me a problem with cdn, but if you're using s3 buckets then to host images then you can't use cloudfalre with it.
Your option will be using cloudfront(aws' CDN).
And yes, in my experience the prices are resonable
@chowdercake this guy knows what he's saying
just don't go 100% spot, have a couple reserved/on-demand instances in your autoscaling group as well... aws have some shit checking for exclusive spot usage and will fake price bursts every now and then to shut you down to force you to reserve, so don't wait to learn your lesson from downtime when it's easy enough to counter with some autoscaling rules.
also you probably will be better off running your own DB rather than paying a premium for RDS which is really useless - if shit hits the fan you'll be stuck on hold like forever until their techs rub one out and eventually decide "your issue is not covered by our SLA"
in fact if you're ready to pay $100+/h you can probably deploy to another provider that's less of an elegant ripoff unless you've got a fat VC behind you?
I learnt it hard way when I found out digital ocean takes 5$ while aws gives same instance for 10$~ (after free tier)
There're new options like t3.nano (2 cores and 0.5 gb ram) but haven't tried it on on-demand yet
And yea, data transfer rates are bitch.
Your cost 5 spot instances will be less than the monthly data transfer rates if you're servers hit traffic
@molaram why is RDS useless, I used in production and I had 0 problems, if I did then it would be my own problem and I would just spin another instance up or fix it?
If you are going to host your own database instances then you might as well host the whole thing, I don’t see how it would be a good idea to split the location of the database and the service? I realise now you might mean host your own db on AWS instead of using rds. I still wouldn’t bother with that.
@molaram the company I work for hosts all our databases in a data centre with our own hardware, when you’ve done this for a while you will realise why services like AWS exists.
I don’t see the benefit of hosting your own database, you will have to do the backup/maintenance plan yourself and it can get messy fast.
In my experience it comes down to the storage strategy and data retention more than anything else. If they don't have true data lake, snapshot+aggregate and a pervasive archival strategy to allow for removal of old data to slower static storage, and have monster clusters where every record since the beginning of time is active indexed state data in second form normalization, AWS will eat their lunch.
Those same companies tend to do things like run reports from application databases and not segregate read stores from query stores, and think Kafka refers to existentialist literature.