Version 7.7 adds a new
apm_remove_span_metadata processor to the default
apm ingest pipeline.
To save storage, this processor removes span metadata fields, like
that are already available on the parent transaction.
While this processor is new, the functionality is not. In previous versions of APM Server, this functionality was hardcoded internally. However, there are use cases for including this metadata in span documents. Switching this metadata cleanup functionality to a processor allows you to define the span metadata that is important in your architecture.
Failing to follow the upgrade steps will result in increased span metadata ingestion when upgrading to version 7.7.
In order to avoid breaking custom pipeline definitions,
pipelines are not overwritten when restarting APM Server.
Enable pipeline overwriting by setting
Copy any custom pipelines you’ve created or altered to the
apmpipeline definition file, located at
- Restart APM Server.
An alternate upgrade method is to use the
This upgrade method can be useful if you’re using an output other than Elasticsearch.
In this case, you can use the command line to disable the active output
and point APM Server at your Elasticsearch cluster.
Then, run the
setup --pipeline command to register the pipeline definitions in
Here’s an example of what this might look like, if you typically output to Logstash:
apm-server setup --pipelines \ -E apm-server.register.ingest.pipeline.overwrite=true \ -E output.logstash.enabled=false \ -E output.elasticsearch.enabled=true \ -E output.elasticsearch.hosts=["http://username:password@elasticsearch:9200"]
setup, restart APM Server.