Day 2 and 3 of Microsoft’s biggest learning conference gave Tom the opportunity to take part in Azure evaluation, dive deeper into some hot-topic services, and make progress bringing Azure Functions and other workloads to Kubernetes that are easy to scale.
Azure Focus Group Feedback
“I’ve been involved in several focus groups on Azure Kubernetes Service, Azure Monitor, Azure Logic Apps, Azure Functions, and Azure Portal, where participants provided their opinion about these services and where we think they can be improved. Here are some key points from the feedback:
- In general – How we can visually represent our resources in Azure, such as Azure Functions, so that new people can better understand what they are doing
- Functions – Having more visibility in how our Azure Functions are scaling with charts and having the ability to receive Event Grid events so that we can react on them. An example is that we’d want to avoid running 1000 instance that were added in last hour, because probably is something wrong and we don’t want to burn our budget.
- AKS – How we can leverage AKS as a more PaaS instead of CPaaS (Cluster PaaS), allowing us to delegate the operational part to MSFT
- Portal – Being able to request a resource in Azure rather than creating it myself. Use case is that at a lot of companies don’t always have the permissions to create them, but we know exactly what we need. Instead of writing an email with all the specs and the correct person creating it for us (if we even know whom to contact), he can just see what we’re requesting and approve the creation.
- Logic Apps – Reducing the Logic App development and making it easier to change them and being able to test them locally. Also, can we get an on-prem solution please? This would be important for companies who are still running BizTalk while having (something similar to) Logic App on-premises. That would be great because not all organizations can migrate because due to limitations with compliance, latency, connectivity, etc.
- ARM – Being able to have deployment retention rather than hitting an 800 deployment maximum and needing to automatically delete the older ones
Investments in Cloud Native Apps
“Brendan Burns gave more insight on the investments Microsoft has made in building cloud-native applications. This is the future. Companies should certainly aim to start all new projects there and use a cloud-based mindset – only the exceptions will not run on the cloud.
Burns shared the story about how he was using Azure Arc to manage a Kubernetes cluster running on his computer and a cluster in his basement. (Fun fact: “Haiku” was the internal codename for Azure Arc.) This showed again that, as I mentioned in my first Ignite Update, Azure Arc is huge. It will mainly focus on container-based solutions, in my opinion.
It’s clear that Microsoft is not only going the container/Kubernetes approach, but they are trying to make it a lot easier to build those applications without having to be a deep technologist. For example, you can use VS Code to visually operate Kubernetes, which makes it a lot easier and accessible. Before you had to use the CLI, which is not for everyone.”
PG Collaboration
“Ignite is always a great opportunity to work in-person with people you can usually only collaborate with long-distance. Here I’ve been working in several on-site product groups, Azure Compute and App Platform in particular.
Instead of collaborating online, Jeff Hollan and I have been working in-person to mature KEDA (Kubernetes-based Event Driven Autoscaling component). This component provides event driven scale for any container running in Kubernetes. I’ve been working with MS on this open-source product since BUILD, and it’s been exciting to see its progress. This closes a very obvious gap and allows us to easily scale our application.
Another reason that makes this progress major is that it supports scaling Azure Functions in Kubernetes, meaning that this allows us to take an Azure Function, which Microsoft used to run, and move it inside our Kubernetes cluster when we need to have more control. This should not be the default, but in certain cases this can certainly be valuable such as on-prem scenarios, compliance, etc.
Currently, we’re focusing our efforts on making the component easier for new people to onboard so that this open-source framework can soon be adopted and released as v1.0.”
That wraps it up for this update. Stay tuned as we continue to post Tom’s highlights throughout the Ignite event!
Subscribe to our RSS feed