[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Data precision
|
From: |
Francisco Reyes |
|
Subject: |
Re: Data precision |
|
Date: |
Mon, 19 May 2014 12:06:03 -0400 |
|
User-agent: |
Mozilla/5.0 (X11; Linux x86_64; rv:24.0) Gecko/20100101 Thunderbird/24.5.0 |
On 05/16/2014 02:57 AM, Martin Pala wrote:
> It will be better to add separate "maximum" based dataset to
> current "average" dataset.
How about keeping any data which generated an event?
For example had apache with 49 process children which was over the 40 we
have an alarm for. That was hours ago... now when I look at the charts
it shows a max of 22 children. That is less than half of the actual max.
If you were to keep event data, then the 49 would be preserved. This way
you only keep data which the person has already indicated is relevant
and meaningful to them because they setup an alarm for it.
Ideally you would copy over the data from the event que to the regular
graphing retention storage since someone could have different retention
for events.