Hi and happy new year to you all!
My first conundrum of 2013... Anyone got any ideas on how I get the amount of time in any given time period that overall FCP read/write latency has been above a certain threshold? The threshold is already set in perf advisor and has been for some time.
I have the usual free opsmanager tools to hand, plus Balance.
I know there are other performance metrics I can and do look at, however for this report I have very explicitly been asked to find something that can be expressed as a percentage, and so far this is the best I can come up with (% of time in a month spent over the threshold). Any other ideas would be welcome though.
Thanks, Peta
On Jan 3, 2013, at 12:20 AM, Peta Thames petathames@gmail.com wrote:
Hi and happy new year to you all!
My first conundrum of 2013... Anyone got any ideas on how I get the amount of time in any given time period that overall FCP read/write latency has been above a certain threshold? The threshold is already set in perf advisor and has been for some time.
Hmm…at my old job I used a SaaS monitoring system called LogicMonitor.
I pulled latency metrics all the time but for NFS, I didn't have any FCP connected systems (besides the filer itself).
If you check out LM - tell him you heard of it from me. I don't get any type of benefit except to let them know where a suggestion comes from.
Thanks for the reference Mike. I saw this post but didn't want to post promotional stuff..
You could indeed do this kind of report with LogicMonitor (and possibly other systems) as an SLA report - set a threshold for latency (20 ms for reads, say) and then run an SLA report to show the amount of time you were/were not in compliance.
On Fri, Jan 4, 2013 at 7:57 AM, Mike Horwath drechsau@gmail.com wrote:
On Jan 3, 2013, at 12:20 AM, Peta Thames petathames@gmail.com wrote:
Hi and happy new year to you all!
My first conundrum of 2013... Anyone got any ideas on how I get the
amount of time in any given time period that overall FCP read/write latency has been above a certain threshold? The threshold is already set in perf advisor and has been for some time.
Hmm…at my old job I used a SaaS monitoring system called LogicMonitor.
I pulled latency metrics all the time but for NFS, I didn't have any FCP connected systems (besides the filer itself).
If you check out LM - tell him you heard of it from me. I don't get any type of benefit except to let them know where a suggestion comes from.
-- Mike Horwath _______________________________________________ Toasters mailing list Toasters@teaparty.net http://www.teaparty.net/mailman/listinfo/toasters
-
Mike Horwath -
Peta Thames -
Steve Francis