I am looking for a way to catch devices where memory usage is slowly crawling up.
We occasionally have outages where the only resolution seems to be to reboot the device, as it is impossible to logon to the device, even if we have out of band management.
In Orion we can see that memory has slowly crawled up to close to 100%. On some devices it may take a week or two and on some devices it can take a year.
I believe a report with a memory usage chart over the last three months would help me catch most of the offending devices by just skimming through the report once a week but I haven't been able to figure out how to create a report for all our Cisco devices with one chart per device.
Has anyone else been in this situation and figured out a way to identify memory leaks?
i guess the same would apply to servers and applications.
Servers, applications, interface utilization spikes.
I gave some example SQL code in this thread:
Re: Help... how can I alert on a sharp change in the WAN Utilization?
No reason that couldn't be modified around to look at memory. Premise is to look at utilization for the last hour/day then compare it to the average of the last week / month. If there is a delta greater than "x" - throw an alert. Pretty low system load, because if you are looking at the last hour vs. last week, no need to run the monitor more than hourly.
SolarWinds uses cookies on its websites to make your online experience easier and better. By using our website, you consent to our use of cookies. For more information on cookies, see our cookie policy.