x
/
fuel-plugin-lma-collector
Code Issues Proposed changes

Reduce the monitoring scope of Rabbitmq queues 

This patch removes the monitoring of all queues individally. This is to
reduce the number of metrics sent by collectd and avoid to reach the low
watermark limit. We just comment the three lines in case we want to
reactive them quickly for debugging purpose.

Change-Id: I9e442c63a47a30e62362f723d51b79ee347a3f8f
Closes-Bug: #1549721
This commit is contained in:
Guillaume Thouvenin 2016-03-25 11:49:51 +01:00
parent 37abff10a4
commit bcd70353f3
2 changed files with 12 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
deployment_scripts/puppet/modules/lma_collector/files/collectd/rabbitmq_info.py
View File

@ -159,9 +159,15 @@ class RabbitMqPlugin(base.Base):
stats['messages'] += ctl_stats[1]
stats['memory'] += ctl_stats[2]
stats['consumers'] += ctl_stats[3]
stats['%s.messages' % queue_name] = ctl_stats[1]
stats['%s.memory' % queue_name] = ctl_stats[2]
stats['%s.consumers' % queue_name] = ctl_stats[3]
# The following three lines are commented out because per-queue
# metrics may result in collectd dropping metrics due to
# over-flooding. See bug
# https://bugs.launchpad.net/lma-toolchain/+bug/1549721
# To reactivate this feature just uncomment them.
#stats['%s.messages' % queue_name] = ctl_stats[1]
#stats['%s.memory' % queue_name] = ctl_stats[2]
#stats['%s.consumers' % queue_name] = ctl_stats[3]
# we need to check if the list of synchronised slaves is
# equal to the list of slaves.
try:

3
doc/source/user/releases.rst
View File

@ -28,6 +28,9 @@ Version 0.8.1
* Avoid spawning several hekad processes (`#1561109
<https://bugs.launchpad.net/lma-toolchain/+bug/1561109>`_).
* Remove the monitoring of individual queue of RabbitMQ (`#1549721
<https://bugs.launchpad.net/lma-toolchain/+bug/1549721>`_).
* Changes
* Upgrade to Heka 0.10.0