Feb. 14, 2019
Every months, we are always getting very interesting set of questions from our customers. We are going to start to release some of them here.
Usually theses questions are not big enough for a full dedicated blog post, but now they found a home!
This month we will talk about .NET, Media Workflow and container management.
Feel free to contact us if you have any comments or recommendations how we should have answered in a better way to our customers. If you would like us to go deeper on one of these topics with a full article. Contact us!
Migrating a lot of Windows applications to AWS. We usually try to get them moved first to .NET core and running on Linux but in some specific case where the application requires a full .NET framework, we have to do a lift & shift.
Usually for people not managing an AWS ECS cluster, we just advise them to run their application on EC2 and potentially to take advantage of some queuing and auto-scaling group to stop the instance when it’s not required to get it up and running.
In the long term, we will help them to build a Windows container of their application and run it on AWS ECS. AWS ECS allows you to take more advantage of the high availability, resources efficiency and elasticity of AWS.
As soon as you start a Fargate task. You will get back an identifier, from this identifier you can do some polling to know what is the status of your task.
One way we advise to our customers is to add inside the running container a call sending a notification the task is completed before to exit.
Need to automatically to quality control some media files, looking for an enterprise solution? You should look at Interra Baton!
Baton is certified on both Windows 2012 R2 or 2016 R2.
We would recommend following types of EC2 for 1 enterprise Baton license:
Baton license is tied with MAC address of an instance. Hence, the EC2 should have Elastic IP configured for license to work.
This month we had an interesting question, where our client wants to get a completion status from his API, except some of the tasks can take up to 24 hours. Our first wrong answer was to recommend to use SQS, but one of their engineer was fast to show us it was not the best way to do it. After some tinkering around this topic, we came with a proof of concept around AWS Step Functions.
AWS Step Functions offers a delay state:
The delay state coupled with some information about the API our customer needs to query will allow them to optimize their query status workflow.
Having be an expert of the AWS cost optimization through TrackIt app for over last 4 years. We are always getting the interesting cost optimization and reporting questions.
How would do it? It’s not an easy task, AWS doesn’t really allow to report and segment the usage of your bandwidth like the way you can do it with EC2 or S3.
The only way who came in our mind so far will be to use VPC flow logs and isolate the different bandwidth workflow we want to monitor in dedicated VPCs.
Definitely not an easy way to answer this customer pain point. Any suggestions or products available in the market to do it in an automated way?
For a lot of our clients Reserved Instances is very hard concept to understand, we had at least four recurring questions coming this month around it. I will only talk about two of them this month and keep the next one for the next post.
First you need to keep in mind reserved instances are just a financial mechanism applied to your AWS bill.
For example if you start a new contract on February 13th (today), it will be effective right away. And if you decide to pay full upfront for 1 year, you will get charged on your AWS bill on February 13th! This charge will cover the next 12 months of your usage for the reserved resources.
Today only Amazon EC2 Standard Reserved Instances can be sold in the Reserved Instance Marketplace. Convertible Reserved Instances cannot be sold. There must be at least one month remaining in the term of the Standard Reserved Instance.
So it’s not really possible to sell any other unused type of resources, but if it happens, open a support ticket with AWS, support can make an exception and help you to cancel them!