The RUM agent sends the queued events to the APM server in a single requests and clears the queue regardless of whether APM server was able to process those events. Making a request to the APM server is the only way the agent can know if the server is able to process the request or not. This should not affect the availability of your application since these events are sent via a separate asynchronous http request.
Would you please explain a bit more, in what way the failing requests affect your application?
The error log information is only there for your information. This error simply means that APM server (or Elasticsearch instance) can not handle the load and requires more resources. Since the APM components function independent from your application this shouldn't affect your application. However, it is recommended to either increase the resources on your servers or adjust the transactionSampleRate in the RUM agent configuration to avoid this error.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.