Ah…I have not explained my use case very well…my apologies:

I am aware of the wildcards and am using them, but say I want have some very 
granular memory tracking for debugging purposes, I might want to have one 
series for each host and port combination.  This is totally feasible if there 
are a small number of workers and/or hosts, but not so much if there are 
hundreds.

The topology I am working on right now is running 128 separate workers, so, in 
order to display granular memory usage, I need to write out the query string 
thusly:

topology.*.--system.host1.6701.%. memory/heap.usedBytes
topology.*.--system.host1.6702.%. memory/heap.usedBytes
topology.*.--system.host1.6703.%. memory/heap.usedBytes
topology.*.--system.host1.6704.%. memory/heap.usedBytes
…
topology.*.--system.host1.6789.%. memory/heap.usedBytes

This is fairly tedious to manually create each series…is there a way to 
parameterize the host and port, for example such that multiple series are 
created automatically?


From: Aravindan Vijayan <[email protected]>
Reply-To: "[email protected]" <[email protected]>
Date: Friday, September 22, 2017 at 3:22 PM
To: "[email protected]" <[email protected]>
Cc: Sriharsha Chintalapani <[email protected]>, Jungtaek Lim 
<[email protected]>
Subject: Re: At my wits end with Ambari Metrics Errors. No Storm and Kafka data 
and unable to start Metrics collector after cleanup

support

Reply via email to