node-red-contrib-kusto

Custom Node-Red node to inject data in Kusto (Azure Data Explorer)

Azure Data Explorer is a high performace timeseries database, query engine and dashboarding tool.

The Node Red Kusto (Azure Data Explorer) integration allows you to send messages directly into events to Kusto also known as Azure Data Explorer for analytics and dashboarding.

Create a free Azure account

• Create a free Azure account. you will be asked for creditcard info, but all rescources created here are free.

Create a Service Principal (App registration)

For Home Assistant to authenticate with Azure Data Explorer, it need a Service Principal

1. Create a Service Principal follow guide step 1-7
2. Copy values for later use:
   • Application (client) ID <--From App registration owerwiev
   • Directory (tenant) ID <--From App registration owerwiev
   • Secret value <--From when the secret was created in 1.7

Create Free Azure Dataexplorer cluster and Database

There are two ways of creating an Azure Data Explorer Cluster: Pay as you go (PAYG) or Free to create a paid cluster follow instructions from here: Microsoft quickstart However Microsoft has released a free offer and this guide describes how to set up a free Azure Data Explorer Cluster and database:

There are a few different between the PAYG and Free versions:

Feature	PAYG Cluster	Free Cluster
Ingestion	Streaming and Queueing	Queueing only (for now)
Cluster size	Fully scalable	4 vCPU, 8GB Menory, ~100GB data

1. Navigate to aka.ms/kustofree.
2. Navigate to My Cluster .
3. And click the Create Cluster button.
4. Name the Cluster and database.
5. Copy the database name for later use
6. Check terms and condition (after reading them) and click Create Cluster.

Within a 30 seconds, you will have a Azure Data Explorer cluster ready

After the creation, copy the Data ingestion URI from the top of the page

Create Azure Data Table

1. Navigate to aka.ms/kustofree.
2. Navigate to Query.
3. Write and execute the foloing statement, replacing <DatabaseName> with the name from cluster creation and <TableName> a name of your choise.

.execute database script <|
// Add SP ingestor rights
.add database ['<DatabaseName>'] ingestors ('aadapp=b5253d02-c8f4-4a79-a0f0-81xxxxxxx2a1f;72f988bf-xxxxx-41af-xxxxx-2d7cd011db47')
// Add SP viewers rights
.add database ['<DatabaseName>'] viewers ('aadapp=b5253d02-c8f4-4a79-a0f0-81xxxxxxx2a1f;72f988bf-xxxxx-41af-xxxxx-2d7cd011db47')
//Alter Table batchin ploicy
.alter database ['<DatabaseName>'] policy ingestionbatching @'{"MaximumBatchingTimeSpan":"00:00:05", "MaximumNumberOfItems": 500, "MaximumRawDataSizeMB": 1024}'
//Drop table if exists
.drop table ['<TableName>'] ifexists
//Create table
.create table ['<TableName>'] (timestamp: datetime, topic: dynamic, payload: dynamic)
//Create JSON mapping
.create table ['<TableName>'] ingestion json mapping 'nodered_json_mapping' '[{"column":"timestamp","path":"$.timestamp"},{"column":"topic","path":"$.topic"},{"column":"payload","path":"$.payload"}]'

If only some af the payload needs to be stored, this example will extract only the needed part as it reaches Azure Data Explorer:

{"column":"part_of_payload","path":"$.payload.key_in_payload"} 

you can extact a part of the json payload and put it in a seperate column.

Example (KNX messages via MQTT and use node-red to store them in ADX):

//Create table
.create table ['knx'] (timestamp: datetime, topic: dynamic, GA: dynamic, val: decimal, knx_textual: dynamic, knx_src_addr: dynamic, knx_dpt: dynamic, payload: dynamic)
//Create JSON mapping
.create table ['knx'] ingestion json mapping 'nodered_json_mapping' '[{"column":"timestamp","path":"$.timestamp"},{"column":"topic","path":"$.topic"},{"column":"GA","path":"$.payload.GA"},{"column":"val","path":"$.payload.val"},{"column":"knx_textual","path":"$.payload.knx_textual"},{"column":"knx_src_addr","path":"$.payload.knx_src_addr"},{"column":"knx_dpt","path":"$.payload.knx_dpt"},{"column":"payload","path":"$.payload"}]'

Configuration

Insert values for Cluster Name and Table Name from previous step. Insert Valuses from the creation of the Azure Service principal created earlier. Insert a value for buffer time (the time NodeRed buffers messages before sending them to Kusto)

if using a free cluster, un-check the Use Streaming API in the form, as this is not supported.

After completiing the flow, Node Red is sending data to Kusto - Azure Data Explorer.

Node Red is buffering for defualt 5 seconds before sending, and Batching Policy in Azure Data Explorer will futher batch up, so expect a litle delay before data is in Kusto.

Using Azure Data Explorer

As the setup is complete, data is being sent to Azure Data Explorer, and you can start exploring your data. Here are som rescources to learn to use Azure Data Explorer

• MS Learn: https://aka.ms/learn.kql, https://aka.ms/learn.adx
• You tube: Official Microsoft Azure Data Explorer YouTube channal
• Blog: Official Microsoft Data Explorer blog