I think you could use the following config options to set the environments
for JobManager and TaskManager.
And then you could use the envs in the log4j configuration file.
"${env:PIPELINE}" could be used in log4j2.containerized.master.env.PIPELINE: my-flink-pipeline containerized.taskmanager.env.PIPELINE: my-flink-pipeline For log4j2, I am afraid you need to set the java dynamic option[1] to get a similar effect. [1]. https://ci.apache.org/projects/flink/flink-docs-master/deployment/config.html#env-java-opts Best, Yang Sidney Feiner <sidney.fei...@startapp.com> 于2020年12月6日周日 下午10:13写道： > Hi, > > We're using Apache Flink 1.9.2 and we've started logging everything as > JSON with log4j (standard log4j1 that comes with Flink). When I say JSON > logging, I just mean that I've formatted in according to: > > log4j.appender.console.layout.ConversionPattern={"level": "%p", "ts": > "%d{ISO8601}", "class": "%c", "line": "%L", "message": "%m"}%n > > > Now I would like to somehow add a field to this JSON to indicate which > pipeline generated the log . At first I thought I'd add another field that > logs some environment variable like such: > > log4j.appender.console.layout.ConversionPattern={"level": "%p", "ts": > "%d{ISO8601}", "class": "%c", "line": "%L", "pipeline: "${PIPELINE}", > "message": "%m"}%n > > But that doesn't seem to be working (is it because the TM is inited before > the pipeline and that's when the placeholders are set?). > > Do you know of a way I could add a field of the current pipeline running? > In my "Main" I have access to the pipeline name and I also have access to > this variable in the tasks themselves. I would prefer not needing to > explicitly using this variable when I log, but that it would be automatic > during logging. > > If anybody has an idea, I'd love to hear it (we can use logback or > anything else if necessary), > > Thanks :) > >
