Given the title indicating it is ‘current,’ it should probably reflect 2017 facts. A couple of corrections to pass along.

SCALING

Heroku not only provides “Manual horizontal and vertical scaling, it also offers:

• Heroku Threshold-based Auto Scaling (https://blog.heroku.com/heroku-autoscaling),
• Heroku Addon-based Auto Scaling: Hirefire (https://hirefire.io)
• Heroku Addon-based Auto Scaling: Adept (https://devcenter.heroku.com/articles/adept-scale)

LANGUAGE SUPPORT

Microsoft .NET Core is available (we’re not talking about Mono, we mean native . NET Core), as are many other Frameworks and Languages, such as R, Erlang, AngularJS, Ember, React etc (all of the JS frameworks), Salesforce Design Experience (Aura JS), Salesforce Lightning Out (Lightning Components wrapped in JS) etc. and Docker (either Docker-to-Heroku or Docker-to-Heroku-Docker-Registry)

AWS & HEROKU SYNERGIES

Not everything is a comparison!

• Heroku offers AWS VPC Peering & VPN capabilities with a Heroku Private Space (secure, isolated stack). This allows the two to work closely, and for customers to use the best environment for the job.
• Heroku Kafka-as-a-Sevice enables a cross-AWS-AZ cluster with a single click. This lets you create a Data Factory, Data Lake or Data Hub that leverages this technology ‘as-a-service’ delivering a simpler, more flexible and cost-effective Data Service Bus (DSB). This, plus Add-ons like Xplenty and Treasure Data, enable AWS/Heroku hybrid cloud environments and take less time to deliver.
• Predictive “Einstein” Heroku services enable you to mine & train predictive applications, whether they live entirely on Heroku, or on AWS, or a combination of both.

I hope this helps expand the dialogue.