AWS Lambda Reviews & Product Details

When Lambda was anounced at AWS re:invent, I heard many people describe it as a game changer. Having now used it for a few tasks, I can confirm this is correct. The price for the value it offers is so incredibly cheap and it's very easy to use.

Supporting multiple languages (we have functions in both node and python) makes it very flexible where anyone could use it from a seasoned software developer to a sysadmin with scripting capability.

Being able to have multiple versions of code and switch between them while running tests on a version is a helpful feature as is the built-in monitoring of executions Vs failures. Review collected by and hosted on G2.com.

Monitoring of failures could be improved. We had to craft a custom solution to take into account the error rate over the number of executions rather than just a flat "notify me when there are 10 errors".

Being able to send logs to an external log service easily would be handy. The built in cloudtrail logs is really lacking compared to tools like loggly and Papertrail. Review collected by and hosted on G2.com.

First & foremost, the simple fact of taking servers/server management out of the mix is invaluable - especially for someone like me, a developer, who doesn't want to get wrapped up in the operations side of things if i don't have to. Review collected by and hosted on G2.com.

Well, there's a little bit of a learning curve in running arbitrary executables with Lambda. You have to launch an actual AWS instance (and a particular one, at that...more or less), and compile the binaries (and possibly related libraries) that your project requires. You then have to include those in all the right places with your packaged up Lambda code.

It would be nice in the future if Lambda expanded the number of available libraries out-of-the-box (e.g., like they do with ImageMagick). Review collected by and hosted on G2.com.

The ability to schedule tasks using SQS. Previously I had to write custom cron jobs in a worker instance and maintaining them has become a pain. With Lambda you know exactly when and what functions will be executed. Review collected by and hosted on G2.com.

It's difficult to integrate the coding to existing CI/CD products.

The startup time/lag can be large at times. This needs to improve. Review collected by and hosted on G2.com.

The integrated monitoring features, the way that lets me assess my application's performance at a fine granularity. Review collected by and hosted on G2.com.

The .zip workflow for uploading functions, lack of support for newer versions of Node, lack of support for languages other than Java, Python, and JS. Review collected by and hosted on G2.com.

Low Cost: it is really awesome that I can setup a reliable scheduling and email messaging service which cost only few cents per month.

Server-less: it is really fun and interesting to shift the programming paradigm to server-less architecture. It does require some effort to get familiar with it, but at the same time, it is some thing really interesting to learn. Review collected by and hosted on G2.com.

I was using python with AWS lambda, the package setup is bit different from the standard python package convention. Deployment is also a bit tricky.

I'd hope Amazon provide better libraries for setting it up as well as deployment. Review collected by and hosted on G2.com.

AWS Lambda is perfect for integrating into an existing application or for creating the backbone of a serverless tech stack. Initially, Lambda costs virtually nothing and the prices are relatively affordable as you scale. I highly recommend it if you're planning to leverage other AWS services like S3, ETS, etc... Review collected by and hosted on G2.com.

The only area it's lacking is in documentation and that can be said for a lot of AWS services. Last time I checked, there wasn't a lot of documentation for getting new users up-to-speed. Fortunately, more developers are adopting it every day, so there are a ton of tutorials online. Review collected by and hosted on G2.com.

AWS Lambda provides me with peace of mind knowing that everything I build on it will be reliable "forever". This allows me to apply it for small projects where maintaining servers would be time prohibitive. Compared to other modern server abstractions like Docker the focus on code over infrastructure requires a refreshingly short learning curve. The potential cost benefit for infrequently used functions is as revolutionary in my option as EC2 was at it's inception. Having the ability to deploy new code without worrying about orchestration, networking, etc, allows me to try new ideas quickly and create rich services with little to no costs for backend services. It really changes the scope of what I can do on small websites with the ability to scale massively under load as needed. Being able to start hundreds of instances of my code for a few seconds allows for the performance of a large EC2 cluster for seconds at a time instead of hours. Review collected by and hosted on G2.com.

As with all Amazon Web Services products working with AWS Lambda does result in some vendor lock-in and requires training that is not very applicable to other systems such as OpenStack. Although in my opinion Amazon is the most innovative and "standard" player in this space the community compared to something like Docker is very small. If AWS Lambda were built on a "standard" like Amazon Elastic Container Service is built on Docker then it would be easier to migrate to another service and apply the knowledge to other platforms. Currently the inability to reserve Lambda runtimes to have cache warmed "instances" limits response times, especially for infrequently called lambdas. The deployment process also requires either a high amount of tooling with something like Serverless, or a slower manual upload through the AWS console. Review collected by and hosted on G2.com.

I really like the ideia of hosting microservices and pay pennies for millons of requests, and it is also fits really well in the AWS product stack. The integration with CloudWatch is amazing, and the versioning is very good too. Review collected by and hosted on G2.com.

I wish it was easier to test and debug them in my own machine, I am currently using the Serverless (http://serverless.com) framework to make this job easier, I think Amazon should deploy their own solution, like Serverless, to make things better. Serverless is very good, but when you have to integrate with other AWS solutions like Cognito, things get a little harder that they should be. Review collected by and hosted on G2.com.

The pricing. This allows me to create code that doesn't run often but I only pay then is used. Usually I need to use Elastic Beanstalk to keep microservices running.

I like that supports node, python and java. I would love to see other options like C#.

I also saw that I can use with Amazon SES to process incoming mails, I haven't had the chance to look at it but its definitely something I'll use. Review collected by and hosted on G2.com.

I still need to master the process of deploying changes. I'll like to have this integrated to git or another solution. Also I would like to have PCI compliance here.

I don´t like that has a request limitation since eventually will prevent us from scale, although it says that that number is increased over the use, it doesn't happen instantly if I need to. Review collected by and hosted on G2.com.

The idea is fantastic - the goal of server-less deployment of code that means no hardware or OS maintenance and the practically unlimited scaling as a result. Review collected by and hosted on G2.com.

Cost/performance ratio is not useful for our use case, which is a shame because on paper it should be a poster-child for Lambda. Review collected by and hosted on G2.com.