Closed hpapazov closed 3 years ago
That's a new one! Can you send over a couple things:
nri-perfmon
(that should be the "source" name). If you find any, please use the "copy" button in them and paste into a text doc.We may also want to run nri-perfmon in verbose mode at command line for a minute and send over its output. That's explained here. I'd say this is a next step though, the current config and log messages are first step.
Our support team did not find anything in the Windows Event Viewer. This is the config:
{ "counterlist": [ { "provider": "Counters", "category": "SMBClientShares", "counters": [ { "counter": "*" } ] } ] }
And this is the command `name: com.newrelic.perfmon description: Perfmon On-Host Integration protocol_version: 1 os: windows commands: metrics: command:
@hpapazov I don't see anything glaring, can you try the latest version of nri-perfmon to see if it happens again? I added some cleanup of old data after each cycle, which may help in case the service is quietly running out of memory.
https://github.com/newrelic/nri-perfmon/releases/tag/0.5.2-alpha
Also can you try verbose mode on one host to see what it outputs? https://github.com/newrelic/nri-perfmon#verbose-logging-mode
Any update @hpapazov ?
Also to note, the latest release is now this one: https://github.com/newrelic/nri-perfmon/releases/tag/0.6.1 I suggest trying it and letting me know if you still see the issue.
Sorry @sschwartzman, our operations team did not have the chance to try the above. For the time being they just disabled it. I am sorry, at this point I do not have any additional information that might be helpful.
Hi, Is it normal to have 20+ nri-perfmon processes running on a given host? Recently we experienced memory issues with couple of our clients' web server hosts. While investigating, our support team discovers 20+ nri-perfmon running there. They told us that it looked like the processes were hung and continually spawned off a new process.
nri-perfmon version: 0.5.1/nri-perfmon-release-x64.zip
Thanks!