Amazon API Gateway vs. Azure Kubernetes Service (AKS)

Experienced a lack of available programming languages while working on a minor project. I had to halt the project and wait for it to be added later. It took ages and had a hit on our productivity. It has a centralized management system which helps and an easy interface which helps to manage multiple tasks in case of large-scale operations and projects.

Incentivized

At one instance, there was a fire in our data center and the backup power had some issues because of which the whole DC went down. I believe with AKS and the replication it's easier to handle such a situation. Also, the scenario would have been pretty transparent to the end-users.

Incentivized

• API Gateway integrates well with AWS Lambda. This allows us to build a web server in the language and framework of our choice, deploy it as a Lambda function, and expose it through API Gateway.
• API Gateway manages API keys. Building rate limiting and request quota features are not trivial (or interesting).
• API Gateway's pricing can be very attractive for services that are accessed infrequently.

Incentivized

• Highly repeatable process
• Minimizes the amount of software to install
• Easily scale resources to meet demand

Incentivized

• Client certificates are troublesome when trying to attach them to API GW stages.
• Debugging across several services can be difficult when API GW is integrated with Route 53 and another service like Lambda or EC2/ELB.
• Creating internal/private APIs, particularly with custom domains, can be unintuitive.

Incentivized

• Need to update cluster manually
• Policies cannot be updated for existing components
• if any change in policy every has to create cluster

Incentivized

We always had a great experience with the AWS support team. They were always on time and very dependable. It was a good partnership while we worked to resolve our issues.

Incentivized

Microsoft support was really good, whenever we raise any ticket they come back to us within a couple of hours.

Incentivized

When we tested Azure API Management at the time, it had serious connectivity issues, it was very unstable, and it needed to do a lot using the command line. Comparing with the AWS solution, which was more mature, and the fact that we have services in use on AWS, we ended up choosing to continue using AWS products. This so as not to run the risk of increasing latency in accesses, and of some functionality not working, due to being developed yet.

Incentivized

The ability to create new instances (i.e. elastic provisioning) is probably the fastest with Azure Kubernetes Service compared to the alternatives that I have looked at. From a pricing perspective, Microsoft always seems to find a way to be more competitive in this area, and literally every other field they play in.

Incentivized

• ROI is negative, you need either to hire them to work with you or spend days/weeks to figure out issues.
• For some of the projects in the end it is not worth it, it is just a "buzz" to use serverless but not practical.
• Service is easy to set up authorization and it is easy to manage.

Incentivized

• Positive impact in moving applications to cloud from the in-hose data center to save cost
• Sometimes got affected with some bugs on the AKS infrastructure but hasn't caused a big kios as our production footprint in AKS is not yet high

Incentivized