Can I avoid requiring static ip?Answered

Thanks for the quick reply.
I would of thought there could of been some kind of alternative, like x-api-key. But oh well. The cost of about $40/M is certainly a bit of a shock!

Thank you for the recommendations, I'll investigate later. 
Are they similar to AWS, or could/should I strip the static IP components from my Architecture and simply route to Linode/Hetzner etc - (I'm trying to find the most time efficient way around this issue)

Are they similar to AWS

It depends on what you're doing. They're a complete alernative to AWS. They have less “bells and wistles”, but also less ways to shoot yourself in the foot (or shoot your bank account in the heart).

I wouldn't say AWS is easier or harder than a VPS (or vice versa), just that the complexity is in different places. AWS forces you to learn their byzantine interfaces and configurations (and, again, hope you didn't spring a leak in your billing). But they also more or less “handle” the configuration of the VMs.

VPS forces you to learn configure your own VMs (so you're learning the underlying system, not some propreitary and evr-changing interface). And they also don't have surprise charges.

So… it depends on how you want to spend your time. Want to learn an Amazon interface that's only applicable to Amazon products (until they decide to change it from under you).

Or do you want to learn the actual operating system you're running, and it's applicable to that operating system no matter if you switch to a different VPS, host internally on your own networks, or just spin-up the OS on your own machine.

Thanks for the detailed reply, it's a lot to think about and I will post my solution when I've settled on one.

My current thoughts:

.VPS does indeed look cost effective, however it's another learning curve considering I'm faily confident with AWS.
.AWS Lightsail seems to offer a solution, but again learning curve
.Countless others have complained regarding the cost of NAT Gateway, one such recommendations is to use ‘fck-nat’ ami, worth a test I guess! 

Minor update:
I'm still working the problem trying all soltions, considering security and minimising risk of the three IP's being suddenly filled. I am considering the following architecture: 

.utilising the echo function once every say 12 hours and scan response for the error (memory tells me this is error 164).
.On detecting error, send a reset trigger from another secure runtime of the three ips (somehow) + email notification to myself.

This line of thinking came about after considering :
.NAT Instance which whilst cheap can still drop out and cause the error without some kind of monitoring
.Having physical hardware like a raspbbery pi server running a staticIP etc just won't do for the scale I'm expecting later on for many reasons - (as much as I'd love to put the spares to use!).
.NATGateway, whilst expensive is indead the least error prone. But still requires this kind of feedback loop incase of outage.

The next question of course, is even semi-automated reset from a secure node even possible?
If not, I'm not sure of a any 98-100% uptime solution.. but it's early days :) 

The next question of course, is even semi-automated reset from a secure node even possible?

No. That would cause LimeLM to block resetting for a period of time.

Just using a static IP address (or a pool of 3 or less addresses) is the only solution.

Apologies for delay, i've finally found a solution with AWS that costs $1.9/ Month. I was ready to move over to Linode/Hetzner/A few alternatives. But I decided to persevere if anything to help others out with similiar use cases.

Here is my related stack overflow post:
 https://stackoverflow.com/questions/76581555/cloudformation-natgateway-replacement-traffic-blocked-from-vpc