S01: E04 – Create a Managed Application

Well-managed applications are critical to delivering efficient cloud operations. The ability to monitor, automate, and add an application’s value into business workflows depend on the operational readiness of the underlying resources. Operational burden on unmanaged resources often lead to security and compliance risks and add to unplanned cost over runs.

Rick: Welcome! I am Rick Hebly and joining me for this OpsTalk™ is Sri Santhanam, Product Manager at MontyCloud. Today, she will show how MontyCloud DAY2™ can help create well-managed applications easily from unmanaged resources and enable customers in the context of their business. Welcome back Sri!

Sri: Thanks for having me Rick. A growing cloud footprint often results in unmanaged resources that are hard to operate. This may be due to several reasons such as shadow IT, missing business context and varying maturity levels of Ops-ready mindset. All this hinders the ability of organizations to realize the full benefits of Cloud adoption. Let me tell you a customer story. A leading online pharmacy’s cloud footprint had 1,000s of resources spread across 16 AWS accounts and multiple regions. They suffered from cost over runs, visibility challenges and ultimately operational inefficiencies. Their unplanned cloud bills spend was growing 30% a year. With MontyCloud DAY2’s well-managed applications feature, we saved them 300,000 annually.

Rick: It is great to see we can help meet customers where they are without causing disruptions. Now tell me, how do you get there? And what is the end-state?

Sri: Of course. In our last Opstalk I demonstrated how DAY2 discovers cloud resources and automatically classifies them into groups that are based on CloudFormation Stacks, Tag Key/Value Pairs and VPCs. Today In four steps, I’ll showcase how to create a well-managed application from one of these groups.

First – I’ll create a server application from set of resources we discover

Second – I’ll set up custom monitoring metrics to later visualize in the dashboard

Third – I will attach automated routine tasks to the application to make it Ops-ready

Fourth – I’ll show any relevant resources that are added later are continuously discovered in the context of this application

Rick: If you haven’t seen the OpsTalk about Discovery & Classification, check it out on this channel. That is awesome. Will this require customers to re-deploy their resources?

Sri: No. Our solution does not require additional provisioning of resources and will not cause disruption to customer’s current workflows.

Rick: Great. Let’s see the demo!

Sri: Sure, Here I am in DAY2 Applications Menu, I am going to now click on the Create application.  Enter the application name, add a description, select the environment.  Now I am selecting “existing resources” as my source and select the Target Account and region.  Using the DAY2 classification by Tags, I am going to apply filter by key values, selecting the Server based Application “Order-Service”. This is a .Net application on EC2 instances with RDS database for the backend behind an application Loadbalancer.

Now, Let’s setup the monitoring for the metrics that are critical for this application. First let me choose the CPU utilization metric for the server instances, same CPU Utilization and Freeable memory, Number of connections for RDS and finally Active Connection Count for the loadbalancer.

Next, we will build the permissions that will be needed to manage this application, I am going to pick from one of these existing permissions say the “Application-Role”.

Finally Let’s review the setup, everything looks good. In just few clicks a well managed application is created from existing resources.

Rick: That click-through wizard makes it really easy and also ensures nothing is overlooked, such as critical metrics. I’d love to see how this app is now a well-managed app that is operations ready.

Sri: Sure!

For that, I will now go to the admin menu to add routine tasks to this application to make it an operations ready server application. Here, we can choose from a list of available automated tasks. For this demo, I am going to attach couple of DAY2 tasks for starting and stopping this application, backup this application with a snapshot task, and enforce compliance of this application with scan/patch operations. Also note, customers can extend this and add their own custom tasks to manage this application. For example, customers can use a script to lookup a business metric from another system (say a Google metric) and automate an operation for this application

As an administrator, I also get to control which of these tasks are available to my users, and I can enable or disable tasks accordingly.

Rick: That is great. Now that the application is operations ready, where can the customer manage it from?

Sri: We do that from our Applications menu, let me choose the same application that we just created. The custom dashboard for this application was automatically generated by DAY2 making it no-code customer ready.

1. Here is the monitoring dashboard for the metrics we configured along with filter by timeline.
2. Resource view to visualize the full inventory of the application according to resource types
3. And tasks view to execute and visualize the task history and scheduled tasks

Rick: It is great to see this well-managed application. Sri, what happens when new resources, say servers are deployed with the same tags/values that we based this application on. Does the customer have to do additional work to bring those resources into this application?

Sri: Great question. With our continuous discovery of resources, customers can achieve this automatically. Let me show you in real-time here. I am going to tag this server with the same tag key value pair that we selected while creating the application. As you can see, DAY2 instantly discovers this new resource, notifies, and classifies it as part of the well-managed application we created.

Rick: Very insightful Sri.

Sri: That’s all I have for today Rick. Thank you for having me again!

Rick: Thank you for contributing to OpsTalk again and no less to the CloudOps community.

It’s clear that many resources are deployed on AWS without application context. They can be orphaned, over-provisioned and undermanaged, leading to all kinds of risks, not least cost-overruns, compliance breaches and security hazards. It’s one thing to get visibility through continuous discovery and classification, its another thing to bring resources back in application contexts and under management. It can be challenging to empowers application administrators to manage routine tasks, and operations team to safely enable self-service operations in the business and application context. That’s the problem I’ve seen solved today: A non-disruptive path from disparate resources to well-managed applications.

Stay tuned for the next OpsTalk. Don’t forget to like and subscribe to the channel.