Metric monitor - Disk IO
Good day,
So I am currently working on full server monitoring and notifications using metrics and alerts.
What I wanted to know if whether anyone has worked out a way to monitor disk IO yet? This is the only outstanding metric that I need for full monitoring of my server.
Thanks in advance for any help :)
-
Hi Stefan,
Do you already have a host metrics source set up on your installed collector on your server? You should be able to collect Disk Reads and Disk Writes as part of this source (look at the tables near the bottom of the link I gave and find the one on Disk Metrics). Is this what you're looking for?
Let me know if this helps!Thanks,
Nick Wilson
Customer Success Manager, Sumo Logic -
Stefan - I am so sorry for the delay here...
I suggest also looking at the host metrics app: https://help.sumologic.com/Send-Data/Applications-and-Other-Data-Sources/Host-Metrics/Host-Metrics-App-Dashboards
This might help you out a bit.
As far as percentages, I'm not incredibly familiar with disk metrics - Would the percentage-based IO be reads / (reads + writes) over a period of time? The Host Metrics App has a dashboard that shows how you would measure read and write rates per host.
Let me know if this still isn't what you're looking for and we can dive a bit further.
Thanks,
Nick Wilson
Customer Success Manager, Sumo Logic
-
Hi Nick,
Apologies I missed this response.
So each disk would have a potential amount of reads and writes it can handle which can be indicated in bytes, but a percentage metric would eliminate the need to tweak the reads and writes metrics every time you create a new type of box or even edit an existing one. I am specifically referring to AWS linux boxes in this case.
Specifically something like this:
https://unix.stackexchange.com/questions/55212/how-can-i-monitor-disk-ioAlso we used to get this metric from Servers using NewRelic, but have decided to move our metric monitoring over to Sumo Logic.
This feature would help a lot!
-
No problem and thanks for taking a look.
In general I think getting percentage on any metrics would be nice for alerting.
Disk usage, Disk IO, Memory (already there), CPU load etc
It just scales way better and gives the option to create a general alert and not have to edit any amounts if you ever change box details.
-
Okay Stefan - I'm still not 100% sure on this one, but one of our sales engineers pointed me to our collectd plugin: https://github.com/SumoLogic/sumologic-collectd-plugin
I think you could set this up on the box to get it to collect that type of metric, but I don't think this metric would be collected natively through Sumo or Cloudwatch.
Do you think collectd would help in your case?
Thanks,
Nick
Customer Success Manager, Sumo Logic
Please sign in to leave a comment.
Comments
7 comments