Interesting one today – standing up a Cosmos DB to record the output of a CycleCloud job run which happened to be written in C++ and started getting “Failed to read item”. Data Explorer stopped showing the results from the item when browsing.
Issue was that our new id had been delimited with slashes and Cosmos DB didn’t like it. If you get “Failed to read item” when clicking through then you might have a character in your document Id that Cosmos doesn’t like.
There are some awesome folks out there who share their hard efforts so the rest of us can have an easier job. A few of these that have been really useful sit around work against the REST APIs of key Azure services.
My days of day in day out development are over so I find a lot of my automation “glue” mashing up deployments relies on PowerShell with the odd bit of CLI. Most is a little bit of scaffolding to deploy ARM templates but occasionally a requirement to work with the data plane of a resource appears and I have to resort to manual config.
ARM Template support for configuring resources is always improving but due to timing this isn’t always possible. Sometimes it is really helpful to understand what is going on, and sometimes the only option is REST.
For the latter I thoroughly recommend POSTMAN if you need to interact, though Azure is also improving native API exploring support. I discovered POSTMAN through an azure Friday video with Steven Lindsay who has some really really useful modules on GitHub. This is really helpful for CosmosDB (Documentdb as it was) and really helped me debug some Gremlin issues.
Next is the PowerShell module for CosmosDB which sits over REST and as well as being an awesome example of the kind is also a really helpful module for checking interactions with CosmosDB.
And finally an excellent post by Michał Pawlikowski on connecting to Azure Data Lake Storage with REST API with PowerShell, showing awesome detective work and publishing a bunch of really useful cmdlets.
Kubernetes and AKS in particular is becoming more and more important to us at work. In our experimental facility we have to stand up varying compute platforms; my main project is examining a specific workload on HPC and part of it needs Kubernetes to support some supporting work.
Then I stumbled across a blog by Chris Johnson . I’ve met Chris (officially a “good guy”) exactly twice in Person; once in 2010 in Berlin at an Ignite Session (when Ignite was a smaller scale effort) for SharePoint 2010 where he presented a session on Microsoft Certified Master, and secondly at Ignite in Orlando last year when I made a point of catching him before he presented a session of the Microsoft Cloud Show with Andrew Connell (also officially a “good guy”) and Julia White (yes, that Julia White).
Anyway, this is one of those posts which is as much for my benefit as yours!