August 8, 2007 at 2:41 pm
Does anyone have any experience with Symantec/Veritas' I3 software? We are considering purchasing this tool.
Your comments please.
August 9, 2007 at 12:59 pm
I've used the product for the past 4 years and have generally been happy with it.
Things I like
Low overhead monitor with no traces. Uses a sampling methodology equivalent to running sp_who, this is different than the approach taken from Idera and Spotlight which uses traces.
Uses Wait Stat analysis this is again different than than other tools and provides per query per login per machine or per login statistics with the ability to drill down to the query level. Again this is different than higher level tools which focus on performance counters. A good analogy is if taking a trip do care how many time your tire rotates (performance counter level) or do you care what are things I'm waiting on for that particular route/care and driver. This is much deeper dive than other tools.
Does the right level of aggregation showing you top queries, number of execution and filtering to the top those that use the most resources, as example a 1 second query run 1,000 times would show up above a 10 second query run once.
Many interesting ways to drill through data:
Resource usage by database for shared servers
Resource usage by machine to determine if web farms are truly load balancing
Since I did an evaluation of different performance tools, I noticed one other company which uses a similar approach from Confio called Ignite it isn't as feature rich as Indepth , but I think it is a better product in the Oracle space and does offer better pricing points so I use Ignite for Oracle over the Indepth product. They have a similar SQL product which may be worth evaluating.
August 10, 2007 at 5:54 am
I haven't used SQL Diag but my understanding is that it is perfmon counter based and uses a SQL traces. When looking at these tools there a few very different approches the vendors have taken:
1-Use polling and wait stat analysis along with some intelligent aggregation. The key with wait stats per query is that there is a direct relationship you aren't left guessing.
2-Put a pretty GUI on Perfmon counters and run a trace on the server. Here you have to try and figure out the correlation between a perfomon counter and a query yourself.
3-The last approach is to sniff TDS packets at the network layer. I looked at a tool called Zero Impact years ago, but couldn't tell you much about it.
August 12, 2007 at 8:05 pm
I was recently shown V5 (? The next full version anyway) of Idera's Diagnostic Manager by the vendor and it looks MUCH better than the previous version.
I was swaying towards Quest's Spotlight because you can see at a glance what is going on but Idera seems to be bridging the gap (as it was a bit of a pain to look at).
Will know more when I get to trial a copy though.
Sorry, I don't know anything about the Symantec tool.
A lack of planning on your part does not constitute an emergency on mine.
August 13, 2007 at 9:23 am
Thank you for your input.
Viewing 6 posts - 1 through 5 (of 5 total)
You must be logged in to reply to this topic. Login to reply