Loading

Upgrade EDOT Cloud Forwarder for Azure

This page describes how to upgrade an existing deployment of EDOT Cloud Forwarder for Azure to a later version.

Upgrading an existing deployment means re-running the deployment command with the later Bicep template. Azure will update the existing infrastructure and application code in place.

  1. Download the new Bicep template

    Download the latest Bicep template from the release URL:

    curl -L https://ela.st/edot-cf-azure-template -o ecf.bicep

    You can verify the version included in the template:

    grep 'param version' ecf.bicep

  2. Re-run the deployment command

    Run the same deployment command you used for the initial installation, pointing to the new template file:

    az deployment group create \
    --resource-group <resource_group_name> \
    --template-file ecf.bicep \
    --parameters \
        otlpEndpoint=<otlp_endpoint> \
        elasticApiKey=<elastic_api_key> \
        eventHubPartitionCount=8 \
        eventHubMessageRetentionInDays=1

    Azure will update the existing resources in the resource group to match the new template.

Not all upgrades can be applied in place. Code-only changes are generally safe and unlikely to introduce breaking changes. However, infrastructure changes may not be compatible with previous versions.

For example, EDOT Cloud Forwarder 0.7.0 migrated from the Consumption hosting plan to the Flex Consumption plan. Azure doesn't support in-place upgrades from Consumption to Flex Consumption for Function Apps.

When an in-place upgrade is impossible, follow these steps:

  1. Deploy a new instance

    Create a new deployment with the updated version in a separate resource group, as described in Deploy EDOT Cloud Forwarder for Azure.

  2. Route traffic to the new deployment

    Update your diagnostic settings or Data Collection Rules to stream telemetry to the new deployment's Event Hubs namespace.

  3. Remove the old deployment

    Once you've confirmed the new deployment is receiving and processing data correctly, remove the old resource group:

    az group delete --name <old_resource_group_name>
    Warning

    If you remove the old resource group, any unprocessed data in the old Event Hubs will be lost.