Greetings, I've tried this lab several times and it's always failing at the same point.
I'm using Chrome, have a VPN connection and the instance is up.
I'm following the lab step by step:
1) Create DataFlow
2) Aggregate & Transform
I Added CPU_Usage and realtimeRecorder objects to the editor. I'm following the lab click by click. In the lab, you see the output data type of the realtimeRecorder as "Table". In this fresh instance of EFM there is no output, in fact it is null. Going further, adding the TableAggregration object you can connect the output of realtimeRecorder to tableAggregation but since the output is null nothing works.
You need to give the column a name in your realtime recorder block (the "name 0" property) before it will start to record into the table.
It seems to me that the errors in this lab are due refreshing Chrome. I have a fresh install of a Win10 Pro that I've used to complete most of the labs in IoT. The labs i'm doing are pretty basic and I'm a CCNA (2019) with about 15 years of Dev experience in .NET and J2EE.
The problem that I see is that if you follow the directions in the lab there are several times where the page needs to refreshed before the result is recognized by the sandbox and reflected back to the user. There is a significant 1-2 minute delay before you can see these changes and this is not mentioned in the lab. As a developer with a decent IDE (Visual Studiio/Eclipse/PyCharm) these changes are immediate. I think that's where the disconnect is.
Maybe it's a Chrome issue, not sure but I'm certain you could not do this lab without refreshing at every step.
Thanks for the details Rob!
I'll forward your feedback to the Sandbox engineering team to investigate.
I've been retrying this lab until I can do it in less than 60 mins which seems reasonable. This is 4th or 5th instance I've provisioned in order to complete that goal. IMO, the Dataflow Editor needs work. Here is step 4 of 4 that I just experienced.
I added notes to the included image Everything I did is in this image is self explanatory.
Thanks for the feedback Rob. We will add it to the enhancements to make the lab steps more clear.
Keep it coming.
I think another reason for the disconnect is that you need to troubleshoot what's NOT happening as you do it. Eventually each person will realize what is not happening and go back and do it again. It took me about 5-6 instances before I could complete the lab fully such that it was working. I really can't see the casual CCNA taking this on without having issues and having a lab partner from Cisco to deal with the little stupid things that I've mentioned. I did finish this lab to completion in under an hour after maybe about a week to figure out the web UI.
Perhaps maybe an Eclipse Snap-in designer would be more responsive to the user design than a web app. I've used Eclipse working with Demandware SaaS back in 2009. It's not anything new and seems like a good fit for this use case.