Description
I have a lot of logs generated by 3rd party libraries in my application, which I do not need to have ingested into Application Insights, and those logs generate significant costs on Application Insights. Currently it is only possible to set global log level threshold, which is not sufficient, because the unwanted logs are either INFO or WARN, so to exclude them I would need to exclude also the logs from my application, which makes this solution useless.
In my current scenario I am running a Scala 2.12 Spark application on Azure Databricks runtime 16.3 as a jar task instrumented with -javaagent (agent version 3.7.2) parameter through spark config and APPLICATIONINSIGHTS_CONNECTION_STRING environment variable, and getting a lot of com.databricks.* and org.apache.spark.* logs. But this feature would be applicable to any other scenario.
I would like to be able to filter/configure logs that are send to Application Insight by the agent on package level.
This could be configured either in the instrumentation.logging section or as a log procesor in similar fashion to the metric-filter processor.