Considerations when changing outputs
Elastic Cloud Serverless projects are autoscaled environments, fully managed by Elastic and available on Elastic Cloud. Availability This functionality is generally available and ready for production usage. Serverless interfaces and procedures might differ from classic Elastic Stack deployments. The Elastic Stack includes Elastic's core products such as Elasticsearch, Kibana, Logstash, and Beats. Availability This functionality is generally available and ready for production usage. Unless stated otherwise on the page, this functionality is available when your Elastic Stack is deployed on Elastic Cloud Hosted, Elastic Cloud Enterprise, Elastic Cloud on Kubernetes, and self-managed environments. This documentation corresponds to the latest patch available for each minor version. If you're not using the latest patch, check the release notes for changes.
Serverless
Generally available
Stack
Generally available
Fleet provides the capability to update your output settings to add new outputs, and then to assign those new outputs to an Elastic Agent policy. However, changing outputs should be done with caution.
When you change the output configuration within a policy applied to one or more agents, there’s a high likelihood of those agents re-ingesting previously processed logs:
- Changing the output will cause the agents to remove and recreate all existing integrations associated with the new output, which as a result of the change receives a new UUID.
- As a consequence of the newly generated output UUID, the agents will retransmit all events and logs they have been configured to collect, since the data registry will be re-created.
In cases when an update to an output is required, it’s generally preferable to update your existing output rather than create a new one.
An example of an update being needed would be when switching from a static IP address to a global load balancer (where both endpoints point to the same underlying cluster). In this type of situation, changing to a new output would result in data being re-collected, while updating the existing output would not.