First off, great work. Thanks for the "Show HN". So a few notes...
First, I need a website that is much more fleshed out than this to get me to buy. But this is a Show HN so that's OK.
How does it handle DNS and what does the UI look like compared to API Gateway? Some screenshots would be nice.
What are the features compared to API Gateway? Does it have authenticators and API keys?
Also, the #1 feature API Gateway is missing that I want is the ability to do rate limiting by IP address instead of API key or account.
As an aside, I'd like to share my setup:
I have two t2.micros in multiple Availability Zones, DNS load balanced and health checked (route53) and have a simple GoLang server that takes the HTTP request and maps it to Lambda Proxy format and calls the AWS Lambda API.
The two micros can handle all the traffic I could reasonable throw at it and cost me almost nothing.
So then you ask, well don't I have servers at that point (which defeats the purpose of "Serverless")? Well, yes, but they are almost 0 config servers (just need the Lambda info) and can be immutable.
t2.micro x 2 = $20 / month
Route53 zone = $2 / month
Route53 health check x 2 = $1 / month
Route53 query pricing w/ latency based load balancing = $1 / million queries
Now, if there was an AMI that set all this up for me in the AWS Marketplace for cost + a few dollars a month I would consider it.
I know the website is not perfect yet (and we're not looking for people to buy yet, maybe that should be clearer?), it would be interesting though what in particular you're missing?
Regarding DNS, I'm a bit unsure what you mean. You point your own domain to our servers via a CNAME. SSL can be enabled as well (via LetsEncrypt), the certificate gets generated automatically via a DNS TXT record or, if you've pointed the domain over to us already, without any effort on your side. Of course you can test your setup before changing any DNS entries by using the *.portal.lambdacult.com subdomain we provide you. Load-balancing is done via ELB, not on the DNS level.
Right now, all Portal is doing is to call your Lambda function and returning its response over HTTP(S) - no rate limiting or any other extra features. Those will come as people (like you) try out the service and request those features, everything is possible :-) It's certainly interesting for us to hear that you consider IP-based rate limiting the #1 missing feature from API Gateway, might be possible soon! ;-)
The AMI idea is certainly interesting, but our goal is to provide a no-brainer service that doesn't need any setup or scaling. I can see this solution work well for small APIs, but don't you want the same flexibility and scalability as you already get with Lambda? Isn't that the great advantage over "traditional" server-based infrastructure? What happens if you get featured on the App Store over night? Also don't underestimate the necessary overhead for monitoring, outages, software updates, EC2 instances being shutdown, etc. - probably easier to manage over a service than over customer-owned EC2 instances.
Regarding screenshots, why don't you just try it out? It's free for now (and hopefully simple), let me know if it takes longer than expected to setup :-)
I'm a huge fan of Lambda and use it extensively at work, but IMO at the point where you need to solve the problems this solves, you should transition out of Lambda and onto something like EC2
Out of curiosity, if this were to solve problems with endpoint pricing and cold-starts, then what other problems do you see that would still force you to move over to EC2? Just wondering, I've used Lambda a bit but never at scale and these are the only two problems that I can foresee.
The other problems aren't technical, stuff like how much you could save by moving off lambda and lock-in (There are enough subtle differences between serverless providers to make migrating not exactly turn-key)
I see Lambda as a springboard for rarely used or short lived applications. I wouldn't build a core product on it
Thanks for reporting this! We had an issue with one of the EC2 instances used behind Portal which led to a 500-response rate within the last 5 hours of around 6% (575 responses were affected in total). The instance has just been replaced.
I'll investigate and hopefully fix the underlying issue within the day.
Thanks for reporting this! We had an issue with one of the EC2 instances used behind Portal which led to a 500-response rate within the last 5 hours of around 6% (575 responses were affected in total). The instance has just been replaced.
I'll investigate and hopefully fix the underlying issue within the day.
First, I need a website that is much more fleshed out than this to get me to buy. But this is a Show HN so that's OK.
How does it handle DNS and what does the UI look like compared to API Gateway? Some screenshots would be nice.
What are the features compared to API Gateway? Does it have authenticators and API keys?
Also, the #1 feature API Gateway is missing that I want is the ability to do rate limiting by IP address instead of API key or account.
As an aside, I'd like to share my setup:
I have two t2.micros in multiple Availability Zones, DNS load balanced and health checked (route53) and have a simple GoLang server that takes the HTTP request and maps it to Lambda Proxy format and calls the AWS Lambda API.
The two micros can handle all the traffic I could reasonable throw at it and cost me almost nothing.
So then you ask, well don't I have servers at that point (which defeats the purpose of "Serverless")? Well, yes, but they are almost 0 config servers (just need the Lambda info) and can be immutable.
t2.micro x 2 = $20 / month Route53 zone = $2 / month Route53 health check x 2 = $1 / month Route53 query pricing w/ latency based load balancing = $1 / million queries
Now, if there was an AMI that set all this up for me in the AWS Marketplace for cost + a few dollars a month I would consider it.