Most of the conversations I've heard about FSA result retention surround 2 topics:
A) capacity tolerance, and
B) usefulness of result set.
Capacity tolerance is simple enough to understand -- how much space are you willing to allocate to the result sets.
When it comes to usefulness of FSA results -- how "old" is "too old" is the question. Some admins will perform weekly comparisons (say Sunday to Sunday). So it's normal for them to keep results for at least a week. Some will run monthly comparisons. Usefulness is very subjective and really depends on how fast your cluster is changing. I've seen FSA results going back a year or two from customers' production clusters.
RobChang-Isilon
136 Posts
672
1
Posted June 9th, 2017 11:00
Hi Mark,
Most of the conversations I've heard about FSA result retention surround 2 topics:
A) capacity tolerance, and
B) usefulness of result set.
Capacity tolerance is simple enough to understand -- how much space are you willing to allocate to the result sets.
When it comes to usefulness of FSA results -- how "old" is "too old" is the question. Some admins will perform weekly comparisons (say Sunday to Sunday). So it's normal for them to keep results for at least a week. Some will run monthly comparisons. Usefulness is very subjective and really depends on how fast your cluster is changing. I've seen FSA results going back a year or two from customers' production clusters.
Hope this helps.