Steef-Jan is a Microsoft MVP, blogger, author, and frequent public speaker. He has over 20 years of experience as a technical lead developer and architect specializing in custom applications, enterprise application integration, Web services, and Microsoft Azure. Steef-Jan has faced integration challenges in various industries, where he has architected, designed, developed, and supported sophisticated and innovative software using many different Microsoft technologies and products. Furthermore, Steef-Jan is a Microsoft Azure UserGroup board member, the Azure Lowlands Event co-organizer, and the Senior Cloud Editor for InfoQ.
Once the integration solutions play, you need some monitoring in place as well and support for it. These integrations usually support critical business processes. It could be the audit process, invoicing, logistics process, etc. Instead of having them on-prem, being with an intermediate, like a BizTalk Server supporting those processes, you now have the more distributed type of solution in the cloud, leveraging Logic Apps functions, Cosmos DB, etc. Sometimes this hasn’t been thought through enough, or they leverage some services like Azure Monitor and Application Insight. Still, holistically, monitoring the process, end to end, has never been given that much thought. Without monitoring, it becomes hard to answer the questions like, “Where am I processing? Am I experiencing issues? Are they being signaled, or at least identify them quickly enough? How can I identify them if I don’t have the information?”. Lack of Monitoring ends up in these challenges, and it is now essential for organizations to adopt Monitoring once the solution is live.
From the get-go! Once you identify what kind of integrations you need and what the components are, you have to look into how I will support a solution once it becomes live. You need to set up some requirements like, “what are the needs? How do you want to set up the end-to-end monitoring?”. In the end, it’s like “People, Process, and Products.” It should be identified from the beginning, and you can set up documentation that includes how the solution is implemented in more real-time details. The first step of implementing Monitoring is setting up something like a first aid kit and identifying what kind of products you want. These requirements help you determine what type of Monitoring your integration needs, a native Azure Monitor, a custom build, or any other tools from the market.
Security, in the end, is the accessibility to your source and target systems. With Azure, it involves Azure AD, App registrations, Managed instances, etc. Security also comes into play for your staff who need to operate and monitor all those solutions because they also need authorization. An example from Steef’s real life is that the average duration has an expiry date if your app registrations are set up to have trust friends with Dynamics 365. If that app registration expires, your solution might fall over, or it doesn’t work anymore because it’s expired. Sometimes it’s hard to identify what happened, and that’s where security and Monitoring are intertwined or related. Security and Monitoring have things in common; both should be taken very seriously. Certificates are interesting because certificates also have an expiry date. Those are the things that you need to monitor as well, and most people who are supporting integration solutions and monitoring them also should be aware of all the security details.