Get data with Fluent Bit

Fluent Bit is an open-source agent that collects logs, metrics, and traces from various sources. It allows you to filter, modify, and aggregate event data before sending it to storage. This article guides you through the process of using Fluent Bit to send data to your KQL database.

This article shows how to ingest data with Fluent Bit.

For a complete list of data connectors, see Data connectors overview.

Prerequisites

• Fluent Bit.
• A workspace with a Microsoft Fabric-enabled capacity.
• A KQL database with ingestion permissions.
• A KQL queryset, which will be referred to later as your query environment.
• Your database ingestion URI to use as the TargetURI value. For more information, see Copy URI.

Create a Microsoft Entra service principal

The Microsoft Entra service principal can be created through the Azure portal or programmatically, as in the following example.

This service principal is the identity used by the connector to write data to your table in Kusto. You grant permissions for this service principal to access Kusto resources.

1. Sign in to your Azure subscription via Azure CLI. Then authenticate in the browser.

   az login
   

2. Choose the subscription to host the principal. This step is needed when you have multiple subscriptions.

   az account set --subscription YOUR_SUBSCRIPTION_GUID
   

3. Create the service principal. In this example, the service principal is called my-service-principal.

   az ad sp create-for-rbac -n "my-service-principal" --role Contributor --scopes /subscriptions/{SubID}
   

4. From the returned JSON data, copy the appId, password, and tenant for future use.

   {
     "appId": "00001111-aaaa-2222-bbbb-3333cccc4444",
     "displayName": "my-service-principal",
     "name": "my-service-principal",
     "password": "00001111-aaaa-2222-bbbb-3333cccc4444",
     "tenant": "00001111-aaaa-2222-bbbb-3333cccc4444"
   }
   

You've created your Microsoft Entra application and service principal.

Create a target table

Fluent Bit forwards logs in JSON format with three properties: log (dynamic), tag (string), and timestamp (datetime).

You can create a table with columns for each of these properties. Alternatively, if you have structured logs, you can create a table with log properties mapped to custom columns. To learn more, select the relevant tab.

Default schema

To create a table for incoming logs from Fluent Bit:

1. Browse to your query environment.

2. Select the database where you'd like to create the table.

3. Run the following .create table command:

   .create table FluentBitLogs (log:dynamic, tag:string, timestamp:datetime)
   

   The incoming JSON properties are automatically mapped into the correct column.

Custom schema

To create a table for incoming structured logs from Fluent Bit:

1. Browse to your query environment.

2. Select the database where you'd like to create the table.

3. Run the .create table command. For example, if your logs contain three fields named myString, myInteger, and myDynamic, you can create a table with the following schema:

   .create table FluentBitLogs (myString:string, myInteger:int, myDynamic: dynamic, timestamp:datetime)
   

4. Create a JSON mapping to map log properties to the appropriate columns. The following command creates a mapping based on the example in the previous step:

   .create-or-alter table FluentBitLogs ingestion json mapping "LogMapping" 
       ```[
       {"column" : "myString", "datatype" : "string", "Properties":{"Path":"$.log.myString"}},
       {"column" : "myInteger", "datatype" : "int", "Properties":{"Path":"$.log.myInteger"}}, 
       {"column" : "myDynamic", "datatype" : "dynamic", "Properties":{"Path":"$.log.myInteger"}}, 
       {"column" : "timestamp", "datatype" : "datetime", "Properties":{"Path":"$.timestamp"}} 
       ]```
   

Grant permissions to the service principal

Grant the service principal from Create a Microsoft Entra service principal database ingestor role permissions to work with the database. For more information, see Examples. Replace the placeholder DatabaseName with the name of the target database and ApplicationID with the AppId value you saved when creating a Microsoft Entra service principal.

.add database <DatabaseName> ingestors ('aadapp=<ApplicationID>;<TenantID>')

Configure Fluent Bit to send logs to your table

To configure Fluent Bit to send logs to your table in Kusto, create a classic mode or YAML mode configuration file with the following output properties:

Field	Description	Required	Default
Name	The pipeline name.		azure_kusto
tenant_id	The tenant ID from Create a Microsoft Entra service principal.	✔️
client_id	The application ID from Create a Microsoft Entra service principal.	✔️
client_secret	The client secret key value (password) from Create a Microsoft Entra service principal.	✔️
managed_identity_client_id	The client ID of the managed identity to use for authentication.	✔️
ingestion_endpoint	Enter the value as described for Ingestion_Endpoint.	✔️
database_name	The name of the database that contains your logs table.	✔️
table_name	The name of the table from Create a target table.	✔️
ingestion_mapping_reference	The name of the ingestion mapping from Create a target table. If you didn't create an ingestion mapping, remove the property from the configuration file.
log_key	Key name of the log content. For instance, log.		log
include_tag_key	If enabled, a tag is appended to output.		On
tag_key	The key name of tag. Ignored if include_tag_key is false.		tag
include_time_key	A timestamp is appended to output, if enabled. Uses the time_key property.		On
time_key	The key name for the timestamp in the log records. Ignored if include_time_key false.		timestamp
ingestion_endpoint_connect_timeout	The connection timeout of various Kusto endpoints in seconds.		60
compression_enabled	Sends compressed HTTP payload (gzip) to Kusto, if enabled.		true
ingestion_resources_refresh_interval	The ingestion resources refresh interval of Kusto endpoint in seconds.
workers	The number of workers to perform flush operations for this output.		0
buffering_enabled	If enabled, buffers data into disk before ingesting into Kusto.		Off
buffer_path	Specifies the location of the directory where the buffered data will be stored if buffering_enabled is On.		/tmp/fluent-bit/azure-kusto/
upload_timeout	Specifies the timeout for uploads if buffering_enabled is On. Files older than this are ingested even if below size limit.		30m
upload_file_size	Specifies the maximum size of a file to be uploaded if buffering_enabled is On.		200MB
azure_kusto_buffer_key	Azure Kusto buffer key to identify plugin instances when buffering_enabled is On. Required for multiple Azure Kusto outputs with buffering.		key
store_dir_limit_size	The maximum size of the directory where buffered data is stored if buffering_enabled is On.		8GB
buffer_file_delete_early	When buffering_enabled is On, whether to delete the buffered file early after successful blob creation.		Off
unify_tag	Creates a single buffer file when buffering_enabled is On.		On
blob_uri_length	Set the length of generated blob URI before ingesting to Kusto.		64
scheduler_max_retries	When buffering_enabled is On, set the maximum number of retries for ingestion using the scheduler.		3
delete_on_max_upload_error	When buffering_enabled is On, whether to delete the buffer file on maximum upload errors.		Off
IO_timeout	Configure the HTTP IO timeout for uploads.		60s

To see an example configuration file, select the relevant tab:

Classic mode

[SERVICE]
    Daemon Off
    Flush 1
    Log_Level trace
    HTTP_Server On
    HTTP_Listen 0.0.0.0
    HTTP_Port 2020
    Health_Check On

[INPUT]
    Name tail
    Path /var/log/containers/*.log
    Tag kube.*
    Mem_Buf_Limit 1MB
    Skip_Long_Lines On
    Refresh_Interval 10

[OUTPUT]
    [OUTPUT]
    Match *
    Name azure_kusto
    Tenant_Id <app_tenant_id>
    Client_Id <app_client_id>
    Client_Secret <app_secret>
    Ingestion_Endpoint https://ingest-<cluster>.<region>.kusto.windows.net
    Database_Name <database_name>
    Table_Name <table_name>
    Ingestion_Mapping_Reference <mapping_name>
    ingestion_endpoint_connect_timeout <ingestion_endpoint_connect_timeout>
    compression_enabled <compression_enabled>
    ingestion_resources_refresh_interval <ingestion_resources_refresh_interval>
    buffering_enabled On
    upload_timeout 2m
    upload_file_size 125M
    azure_kusto_buffer_key kusto1
    buffer_file_delete_early Off
    unify_tag On
    buffer_dir /var/log/
    store_dir_limit_size 16GB
    blob_uri_length 128
    scheduler_max_retries 3
    delete_on_max_upload_error Off
    io_timeout 60s

YAML mode

config:
  service: |
    [SERVICE]
        Daemon Off
        Flush 1
        Log_Level trace
        HTTP_Server On
        HTTP_Listen 0.0.0.0
        HTTP_Port 2020
        Health_Check On
        
  inputs: |
    [INPUT]
        Name tail
        Path /var/log/containers/*.log
        multiline.parser docker, cri
        Tag kube.*
        Mem_Buf_Limit 1MB
        Skip_Long_Lines On
        Refresh_Interval 10

  filters: |
    [FILTER]
        Name kubernetes
        Match kube.*
        Merge_Log On
        Merge_Log_key log_processed
        K8S-Logging.Parser On
        K8S-Logging.Exclude Off


  outputs: |
    [OUTPUT]
        [OUTPUT]
    Match *
    Name azure_kusto
    Tenant_Id <app_tenant_id>
    Client_Id <app_client_id>
    Client_Secret <app_secret>
    Ingestion_Endpoint https://ingest-<cluster>.<region>.kusto.windows.net
    Database_Name <database_name>
    Table_Name <table_name>
    Ingestion_Mapping_Reference <mapping_name>
    ingestion_endpoint_connect_timeout <ingestion_endpoint_connect_timeout>
    compression_enabled <compression_enabled>
    ingestion_resources_refresh_interval <ingestion_resources_refresh_interval>
    buffering_enabled On
    upload_timeout 2m
    upload_file_size 125M
    azure_kusto_buffer_key kusto1
    buffer_file_delete_early Off
    unify_tag On
    buffer_dir /var/log/
    store_dir_limit_size 16GB
    blob_uri_length 128
    scheduler_max_retries 3
    delete_on_max_upload_error Off
    io_timeout 60s

Confirm data ingestion

1. Once data arrives in the table, confirm the transfer of data, by checking the row count:

   FluentBitLogs
   | count
   

2. To view a sample of log data, run the following query:

   FluentBitLogs
   | take 100
   

Related content

• Query data in a KQL queryset