Code0x58 commented on issue #1453:
URL:
https://github.com/apache/incubator-heron/issues/1453#issuecomment-658918932
I think this would only apply to the the topologies in the executors, which
are written in their own language and user execution environment - generally
this is probably best left to the users to find a balance between their
language, cluster, log shipping, and application logging needs. Even going
through every piece of code that write a log message and may be able to
determine it's topology name, then putting that into the log message in a
consistent format, might not yield much.
That would leave me tempted to close this as a non-issue/no great solution
or requirement.
_p.s. in Kubernetes the general approach is to log to stdout and let leave
aggregation to the cluster's log shipping. For me, the ideal is to have logs
rendered into a JSON envelope (e.g. `{"level": "info", "message": "become the
log", "tag": "value"}`). Then it can be nicely loaded into an ELK stack with a
JSON filter (or similar elsewhere), then elements of data are very easy to
explore and aggregate/filter._
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
