Skip to content

Only prefix metric records with instance name

Description

This MR fixes #180 (closed) by only prefixing MetricRecords with the instance names. LogRecords (appropriately) don't have a name, so this caused the entire monitoring bus to fail when it received a LogRecord. I also removed the unused send_record_nowait method so I could put all of the instance name stuff into a separate helper method.

As a sidenote, this took much longer to find than I expected it to because the monitoring bus didn't output anything when it raised an exception. Does anyone know how to get that output to show up? I suspect the fact that this is running asynchronously is masking it somehow, but I'm not familiar enough with asynchronous programming in Python to know why.

Changes proposed in this merge request:

  • Only prefix metric records
  • Remove the unused synchronous publishing method

This merge request, when merged, will address issue/bug: #180 (closed)

Edited by Rohit Kothur

Merge request reports