Install the Azure Operator Insights ingestion agent and configure it to upload data

When you follow this article, you set up an Azure Operator Insights ingestion agent on a virtual machine (VM) in your network and configure it to upload data to a Data Product. This ingestion agent supports uploading:

  • Files stored on an SFTP server.
  • Affirmed Mobile Content Cloud (MCC) Event Data Record (EDR) data streams.

For an overview of ingestion agents, see Ingestion agent overview.


From the documentation for your Data Product, obtain the:

  • Specifications for the VM on which you plan to install the VM agent.
  • Sample configuration for the ingestion agent.

VM security recommendations

The VM used for the ingestion agent should be set up following best practice for security. We recommend the following actions:


When using an Azure VM:

  • Give the VM a private IP address.
  • Configure a Network Security Group (NSG) to only allow network traffic on the ports that are required to run the agent and maintain the VM.
  • Beyond this, network configuration depends on whether restricted access is set up on the data product (whether you're using service endpoints to access the Data product's input storage account). Some networking configuration might incur extra cost, such as an Azure virtual network between the VM and the Data Product's input storage account.

When using an on-premises VM:

  • Configure a firewall to only allow network traffic on the ports that are required to run the agent and maintain the VM.

Disk encryption

Ensure Azure disk encryption is enabled (this is the default when you create the VM).

OS version

  • Keep the OS version up-to-date to avoid known vulnerabilities.
  • Configure the VM to periodically check for missing system updates.


Limit access to the VM to a minimal set of users. Configure audit logging on the VM - for example, using the Linux audit package - to record sign-in attempts and actions taken by logged-in users.

We recommend that you restrict the following:

  • Admin access to the VM (for example, to stop/start/install the ingestion agent).
  • Access to the directory where the logs are stored: /var/log/az-aoi-ingestion/.
  • Access to the managed identity or certificate and private key for the service principal that you create during this procedure.
  • Access to the directory for secrets that you create on the VM during this procedure.

Microsoft Defender for Cloud

When using an Azure VM, also follow all recommendations from Microsoft Defender for Cloud. You can find these recommendations in the portal by navigating to the VM, then selecting Security.

Set up authentication to Azure

The ingestion agent must be able to authenticate with the Azure Key Vault created by the Data Product to retrieve storage credentials. The method of authentication can either be:

  • Service principal with certificate credential. This must be used if the ingestion agent is running outside of Azure, such as an on-premises network.
  • Managed identity. If the ingestion agent is running on an Azure VM, we recommend this method. It doesn't require handling any credentials (unlike a service principal).


You may need a Microsoft Entra tenant administrator in your organization to perform this setup for you.

Use a managed identity for authentication

If the ingestion agent is running in Azure, we recommend managed identities. For more detailed information, see the overview of managed identities.


Ingestion agents on Azure VMs support both system-assigned and user-assigned managed identities. For multiple agents, a user-assigned managed identity is simpler because you can authorise the identity to the Data Product Key Vault for all VMs running the agent.

  1. Create or obtain a user-assigned managed identity, follow the instructions in Manage user-assigned managed identities. If you plan to use a system-assigned managed identity, don't create a user-assigned managed identity.
  2. Follow the instructions in Configure managed identities for Azure resources on a VM using the Azure portal according to the type of managed identity being used.
  3. Note the Object ID of the managed identity. This is a UUID of the form xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx, where each character is a hexadecimal digit.

You can now grant permissions for the Data Product Key Vault.

Use a service principal for authentication

If the ingestion agent is running outside of Azure, such as an on-premises network then you cannot use managed identities and must instead authenticate to the Data Product Key Vault using a service principal with a certificate credential. Each agent must also have a copy of the certificate stored on the virtual machine.

Create a service principal

  1. Create or obtain a Microsoft Entra ID service principal. Follow the instructions detailed in Create a Microsoft Entra app and service principal in the portal. Leave the Redirect URI field empty.
  2. Note the Application (client) ID, and your Microsoft Entra Directory (tenant) ID (these IDs are UUIDs of the form xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx, where each character is a hexadecimal digit).

Prepare certificates for the service principal

The ingestion agent only supports certificate credentials for service principals. It's up to you whether you use the same certificate and key for each VM, or use a unique certificate and key for each. Using a certificate per VM provides better security and has a smaller impact if a key is leaked or the certificate expires. However, this method adds a higher maintainability and operational complexity.

  1. Obtain one or more certificates. We strongly recommend using trusted certificates from a certificate authority. Certificates can be generated from Azure Key Vault: see Set and retrieve a certificate from Key Vault using Azure portal. Doing so allows you to configure expiry alerting and gives you time to regenerate new certificates and apply them to your ingestion agents before they expire. Once a certificate expires, the agent is unable to authenticate to Azure and no longer uploads data. For details of this approach see Renew your Azure Key Vault certificates. If you choose to use Azure Key Vault then:
    • This Azure Key Vault must be a different instance to the Data Product Key Vault, either one you already control, or a new one.
    • You need the 'Key Vault Certificates Officer' role on this Azure Key Vault in order to add the certificate to the Key Vault. See Assign Azure roles using the Azure portal for details of how to assign roles in Azure.
  2. Add the certificate or certificates as credentials to your service principal, following Create a Microsoft Entra app and service principal in the portal.
  3. Ensure the certificates are available in PKCS#12 (P12) format, with no passphrase protecting them.
    • If the certificate is stored in an Azure Key Vault, download the certificate in the PFX format. PFX is identical to P12.
    • On Linux, you can convert a certificate and private key using OpenSSL. When prompted for an export password, press Enter to supply an empty passphrase. This can then be stored in an Azure Key Vault as outlined in step 1.
    openssl pkcs12 -nodes -export -in <certificate.pem> -inkey <key.pem> -out <certificate.p12>


The P12 file must not be protected with a passphrase.

  1. Validate your P12 file. This displays information about the P12 file including the certificate and private key.

    openssl pkcs12 -nodes -in <certificate.p12> -info
  2. Ensure the P12 file is base64 encoded. On Linux, you can base64 encode a P12 certificate by using the base64 command.

    base64 -w 0 <certificate.p12> > <base64-encoded-certificate.p12>

Grant permissions for the Data Product Key Vault

  1. Find the Azure Key Vault that holds the storage credentials for the input storage account. This Key Vault is in a resource group named <data-product-name>-HostedResources-<unique-id>.
  2. Grant your managed identity or service principal the 'Key Vault Secrets User' role on this Key Vault. You need Owner level permissions on your Azure subscription. See Assign Azure roles using the Azure portal for details of how to assign roles in Azure.
  3. Note the name of the Key Vault.

Prepare the SFTP server

This section is only required for the SFTP pull source.

On the SFTP server:

  1. Ensure port 22/TCP to the VM is open.
  2. Create a new user, or determine an existing user on the SFTP server that the ingestion agent should use to connect to the SFTP server.
  3. Determine the authentication method that the ingestion agent should use to connect to the SFTP server. The agent supports:
    • Password authentication
    • SSH key authentication
  4. Configure the SFTP server to remove files after a period of time (a retention period). Ensure the retention period is long enough that the agent should process the files before the SFTP server deletes them. The example configuration file contains configuration for checking for new files every five minutes.


Your SFTP server must remove files after a suitable retention period so that it does not run out of disk space. The ingestion agent does not remove files automatically.

A shorter retention time reduces disk usage, increases the speed of the agent and reduces the risk of duplicate uploads. However, a shorter retention period increases the risk that data is lost if data cannot be retrieved by the agent or uploaded to Azure Operator Insights.

Prepare the VMs

Repeat these steps for each VM onto which you want to install the agent.

  1. Ensure you have an SSH session open to the VM, and that you have sudo permissions.

  2. Install systemd, logrotate, and zip on the VM, if not already present. For example:

    sudo dnf install systemd logrotate zip
  3. If you're using a service principal, copy the base64-encoded P12 certificate (created in the Prepare certificates step) to the VM, in a location accessible to the ingestion agent.

  4. Configure the agent VM based on the type of ingestion source.

    1. Verify that the VM has the following ports open. These ports must be open both in cloud network security groups and in any firewall running on the VM itself (such as firewalld or iptables).
      • Port 443/TCP outbound to Azure
      • Port 22/TCP outbound to the SFTP server
    2. Create a directory to use for storing secrets for the agent. We call this directory the secrets directory. Note its path.
    3. Create a file in the secrets directory containing password or private SSH key for the SFTP server.
      • The file must not have a file extension.
      • Choose an appropriate name for this file, and note it for later. This name is referenced in the agent configuration.
      • The file must contain only the secret value (password or SSH key), with no extra whitespace.
    4. If you're using an SSH key that has a passphrase to authenticate, use the same method to create a separate file that contains the passphrase.
    5. Ensure the SFTP server's public SSH key is listed on the VM's global known_hosts file located at /etc/ssh/ssh_known_hosts.


    Use the Linux command ssh-keyscan to add a server's SSH public key to a VM's known_hosts file manually. For example, ssh-keyscan -H <server-ip> | sudo tee -a /etc/ssh/ssh_known_hosts.

Ensure that the VM can resolve Microsoft hostnames

Check that the VM can resolve public hostnames to IP addresses. For example, open an SSH session and use dig to check that the VM can resolve to an IP address.

If the VM can't use DNS to resolve public Microsoft hostnames to IP addresses, map the required hostnames to IP addresses. Return to this procedure when you have finished the configuration.

Install the agent software

The agent software package is hosted on the "Linux software repository for Microsoft products" at

The name of the ingestion agent package is az-aoi-ingestion.

To download and install a package from the software repository, follow the relevant steps for your VM's Linux distribution in How to install Microsoft software packages using the Linux Repository.

For example, if you are installing on a VM running Red Hat Enterprise Linux (RHEL) 8, follow the instructions under the Red Hat-based Linux distributions heading, substituting the following parameters:

  • distribution: rhel
  • version: 8
  • package-name: az-aoi-ingestion

Configure the agent software

The configuration you need is specific to the type of source and your Data Product. Ensure you have access to your Data Product's documentation to see the required values. For example:

  1. Connect to the VM over SSH.

  2. Change to the configuration directory.

    cd /etc/az-aoi-ingestion
  3. Make a copy of the default configuration file.

    sudo cp example_config.yaml config.yaml
  4. Set the agent_id field to a unique identifier for the agent instance – for example london-sftp-1. This name becomes searchable metadata in Operator Insights for all data ingested by this agent. Reserved URL characters must be percent-encoded.

  5. Configure the secret_providers section.

    SFTP sources require two types of secret providers.

    • A secret provider of type key_vault, which contains details required to connect to the Data Product's Azure Key Vault and allow connection to the Data Product's input storage account.
    • A secret provider of type file_system, which specifies a directory on the VM for storing credentials for connecting to an SFTP server.
    1. For the secret provider with type key_vault and name data_product_keyvault, set the following fields.
    2. For the secret provider with type file_system and name local_file_system, set the following fields.
      • secrets_directory to the absolute path to the secrets directory on the agent VM, which was created in the Prepare the VMs step.

    You can add more secret providers (for example, if you want to upload to multiple Data Products) or change the names of the default secret providers.

  6. Configure the pipelines section using the example configuration and your Data Product's documentation. Each pipeline has three configuration sections.

    • id. The ID identifies the pipeline and must not be the same as any other pipeline ID for this ingestion agent. Any URL reserved characters must be percent-encoded. Refer to your Data Product's documentation for any recommendations.

    • source. Source configuration controls which files are ingested. You can configure multiple sources.

      Delete all pipelines in the example except the contoso-logs example, which contains sftp_pull source configuration.

      Update the example to meet your requirements. The following fields are required for each source.

      • host: the hostname or IP address of the SFTP server.
      • filtering.base_path: the path to a folder on the SFTP server that files will be uploaded to Azure Operator Insights from.
      • known_hosts_file: the path on the VM to the global known_hosts file, located at /etc/ssh/ssh_known_hosts. This file should contain the public SSH keys of the SFTP host server as outlined in Prepare the VMs.
      • user: the name of the user on the SFTP server that the agent should use to connect.
      • Depending on the method of authentication you chose in Prepare the VMs, set either password or private_key.
        • For password authentication, set secret_name to the name of the file containing the password in the secrets_directory folder.
        • For SSH key authentication, set key_secret to the name of the file containing the SSH key in the secrets_directory folder. If the private key is protected with a passphrase, set passphrase_secret_name to the name of the file containing the passphrase in the secrets_directory folder.

      For required or recommended values for other fields, refer to the documentation for your Data Product.


      The agent supports additional optional configuration for the following:

      • Specifying a pattern of files in the base_path folder which will be uploaded (by default all files in the folder are uploaded).
      • Specifying a pattern of files in the base_path folder which should not be uploaded.
      • A time and date before which files in the base_path folder will not be uploaded.
      • How often the ingestion agent uploads files (the value provided in the example configuration file corresponds to every hour).
      • A settling time, which is a time period after a file is last modified that the agent will wait before it is uploaded (the value provided in the example configuration file is 5 minutes).

      For more information about these configuration options, see Configuration reference for Azure Operator Insights ingestion agent.

    • sink. Sink configuration controls uploading data to the Data Product's input storage account.

      • In the sas_token section, set the secret_provider to the appropriate key_vault secret provider for the Data Product, or use the default data_product_keyvault if you used the default name earlier. Leave secret_name unchanged.
      • Refer to your Data Product's documentation for information on required values for other parameters.


        The container_name field must be set exactly as specified by your Data Product's documentation.

Start the agent software

  1. Start the agent.
    sudo systemctl start az-aoi-ingestion
  2. Check that the agent is running.
    sudo systemctl status az-aoi-ingestion
    1. If you see any status other than active (running), look at the logs as described in Monitor and troubleshoot ingestion agents for Azure Operator Insights to understand the error. It's likely that some configuration is incorrect.
    2. Once you resolve the issue,  attempt to start the agent again.
    3. If issues persist, raise a support ticket.
  3. Once the agent is running, ensure it starts automatically after reboot.
    sudo systemctl enable az-aoi-ingestion.service

Learn how to: