I can't seem to locate the documentation that tells what is indicated when a pipe turns orange, I had assumed missed SLA but mine are changing when 0 SLA
The WD pipe in Workload Monitor should only change color (goes red) when an SLG has been missed in the viewed timeframe. If you hover over the colored pipe does it indicate that "0% requests missed SLG"?
Hmmm. What version of Viewpoint are you running with? I would ask for the full four digit version identifier.
I've opened a request to be upgraded to .06 to address a query spotlight issue that was fixed in this release.
In the Workload health portlet, the same workload is indicated as having missed SLG(2% of requests) and is red.
In Designer the corresponding pipe is red but when hovering over it says 0% have missed SLG.
Hello Everyone...I just need to clear a few things about Workload. Recently, I read about Resource Partition, Performance Group, Performance Period and Allocation Group. From what I have understood, the Performance Group is decided by the User's Account String. And the Allocation group is decided by Performance Period (Milestones), with different Performance Periods alligning to different Allocation Group.
The hierarchy is: Resource Partition -> Performance Group -> Performance Period -> Allocation Group
My question is:
(a) How a query is allocated to a Resource Partition ? (My guess is since the Performance Group is implicitly decided by Account String, a set of Performance Periods constitutes a Resource Partition. Based upon Performance Period, the Resourse Partition is decided).
(b) Once within a Performance Group, based upon milestones, a query is moved between different Performance Period and hence, Allocation Group. My Question is which is the First Performance Period within a Performance Group that is allocated ? Is it something default ?
(c) In all these above scenarios, where does Workload comes into play. Is Workload equivalent to Performance Period ? Where exactly Workload fits in the above hierarchical diagram.
Thanks in advance.
This question was also posted on my blog on Dev X. Here is the response I gave at that location:
A query is not allocated to a resource partition directly. It is mapped to a performance group. A query is under the control of an allocation group, based on which allocation group its performance group (PG) points to. The resource partition is a collection of performance groups. So whatever PG a query belongs to will dictate the resource partition it is associated with.
A performance period is simply an attribute of a performance group. It allows a query to map to a single performance group, but based on time or CPU usage to be moved under the control of a different allocation group than the default allocation group. The performance period has no independent existence outside a performance group. It is simply a bridge between a performance group and an allocation group. It is a carry-over from previous releases where it had a more important role. Today, you can basically ignore performance periods.
Most performance groups only have a single (default) performance period. For appliance platforms, the performance groups (most of them) have built-in performance periods in order to control the automatic demotions. The first PP is the default one where the query starts; the second PP is for the allocation group after the demotion. However, you don't have to be concerned about performance periods at all or how to define them. In fact, on the appliance, you will not be allowed to add or delete performance periods. If you are using full TASM, you don't need to worry about them either, as you have things like workload exceptions to provide similar functionality.
The TASM workload is a performance group under the covers (at least in the SLES 10 versions). TASM workloads do not know about or get involved in performance periods. So if you are a TASM user, you don't have to be concerned or ever think about performance periods. Just set up workload exceptions if you need them, using whatever exception criteria you like.