Azure Time Series Insights is a new service that makes it very easy to store and visualize time series data. In this blog post, we will create a dashboard that looks like the one below (click to enlarge):
The dashboard has four sections:
- Query1: a heat map of events per device; in this case there are 20 devices sending data every 2 seconds
- Query2: a line graph with random “temperature” data
- Query3: a line graph with both “temperature” and “humidity” data
- Query4: a line graph with “humidity” data
The events are sent to an IoT Hub using the following JSON shape: {temperature: x, humidity: y} where x and y are randomized floating point numbers, generated by an IoT device simulator.
Step 1: Create IoT Hub
Install Azure CLI 2.0, and then use az login to login. Use az account list to list your subscriptions and use az account set –subscription name_or_id to set the default subscription. Next, issue the following commands to create a resource group and an IoT Hub (set location to your preference):
|
1 2 |
az group create --name resource_group_name --location westeurope az iot hub create --sku F1 --name iot_hub_name --resource-group resource_group_name |
As a best practice, create a separate consumer group on the Events endpoint. In the Azure Portal, in the properties of the IoT Hub, click Endpoints. Then click Events and add a consumer group underneath $Default. Click Save.
Record the Connection String – primary key setting of the device or iothubowner Shared access policy. Click Shared Access Policies, and device to find this connection string. It will be in the form of:
|
1 |
HostName=iot_hub_name.azure-devices.net;SharedAccessKeyName=keyname;SharedAccessKey=b5dARuGPhL6wdgHboUIhEC6LlcFalIjfEdh4aXYa1WI= |
You will need this connection string later to configure the IoT Simulator.
Step 2: Create Time Series Insights Environment
In the Azure Portal, click the green + and navigate to Internet of Things. Click Time Series Insights and follow the on-screen instructions. You will end up with:
I selected one unit of the S1 tier which is more than enough for this example.
Step 3: Set Data Access Policy
Even though you created the Time Series Insights Environment, you still need to grant yourself access to the data. Click Data Access Policies and add your user or group and a role of Contributor.
Step 4: Add Event Source
We will add the IoT Hub we created earlier as an event source. Click Event Sources and then click Add. Give the event source a name and set the source to IoT Hub. Then select an IoT Hub from your available subscriptions and do not forget to set the consumer group to the one you created in step 1. If your event data has a timestamp, you can enter the timestamp property name. If you do not specify the timestamp, the event enqueue time set by the IoT Hub will be used.
Note that Azure Time Series Insights also supports Event Hubs as an event source.
Step 5: Configure the IoT simulator
Head over to https://github.com/gbaeke/iot-simulator/releases/tag/v0.3 and download iot-simulator.exe to a folder of your choice. In the same folder add a file called config.json with the following contents:
|
1 2 3 4 5 6 7 8 |
{ "Interval":5, "IoTHubs":["iot_hub_name.azure-devices.net”], "SasTokens":["SharedAccessSignature sr=..."], "DevGroups":[ {"Prefix":"ts","DeviceNum":20,"Firmware":"1.0","IoTHub": 0} ] } |
In the SasTokens array, replace SharedAccessSignature sr=… with a Sas token that has the necessary rights to submit events to the IoT Hub. One way of doing so, is with Device Explorer. Once installed, copy the connection string from step 1 in the connection string box and click Generate SAS. Copy the Sas token in the config.json file.
With the config.json correctly configured, from a command prompt, start iot-simulator.exe. It will connect to the IoT Hub, create the devices and start sending data every 5 seconds from every device. In the sample config file, you can set the interval in seconds (Interval) and the amount of devices (DeviceNum). To clean up the devices, run iot-simulator.exe –r.
Step 6: Visualize the data
Now go to https://insights.timeseries.azure.com and login with the credentials you used in step 3. You will get a screen to select data. I selected Last 60 Mins from the quick times dropdown and then clicked the search icon:
In the following screen, click Heatmap and then configure the box at the left with a descriptive title. Also select a split by deviceid to have an idea about the number of events per time window per device and to spot devices that stopped sending data.
Now, at the right top corner, click the circle with the four squares. You end up with:
Now click the + in the top, right section. Select a time range again and then, at the left, change the measure from Events to Temperature. Automatically, the temperature will be averaged over the interval size. Change the term (Term 1) to Temperature and click the circle with the four squares again.
The temperature line graph has been added and you can now click the copy icon and create the same visualization for humidity.
Now it’s easy to create the other panel with both temperature and humidity. Give it a go and try out other visualizations. When you are finished, you can click the Save icon and save this perspective. Yep, these visualizations are called perspectives!
It’s still early days for the service and many features will be added in the near future. If you are already working with event data coming into an Event Hub and IoT Hub, it should be easy to add a new consumer group and start analyzing the data with this service.