Comments
-
Unfortunately we don't have an ad hoc report rendering engine for DPA. If the report doesn't exist in our dropdown canned list, then we're talking about basically just running a query against our repo. Now having said that, there are some cool reporting tools out there that can create the report based on the data pulled…
-
Thinking you will have to provide logs for us. Can you open a support case on this one? Submit a Ticket | SolarWinds Customer Portal
-
If you set the interval for the alert to check every 3 hours, I believe it averages the individual collections for that metric and should achieve what you want. You should test this out though to make sure it's the desired behavior.
-
That was thought of, but many companies have a shared login to DPA. We could however auto-populate, but still have the field be editable for cases where shared logins are used. Great to hear feedback like this though as the product managers love it!!
-
Mike, over 90% of the storage consumption is held in the detailed data (most recent 30 days). So, even if you removed the summary data older than 6 months, the impact/benefit to storage would be minimal. There is an option called CLEAN_DAYS_OF_DETAIL that allows you to store less than the 30 days of detail (what you see in…
-
glimbeek, we'll look into this and see what we can come up with. You're right that the code in that query is for Oracle - should have been under custom Oracle reports... We'll get that moved over also.
-
Just to be clear, the index recommendations you are seeing are coming from SQL Server. Here is a good article on it. Find missing indexes using SQL Servers index related DMVs | Basit's SQL Server Tips We're just reporting what we find from these DMVs (just like all tools out there currently). I think your question would be…
-
Yeah, what you are looking at within each dimension is the top 15 for the timeframe you are in (so 30 days, 1 day, 1 hour, etc.). So on SQL, it's the top 15 and their relative wait time. For Waits, it's the top 15 waits and their relative waits. etc. They will differ from dimension to dimension. Likely the closest to total…
-
It looks like your question is more targeted for Orion rather than DPA. I'd suggest posting this again under SAM or NPM or one the Orion core products.
-
This is currently a feature request for the DPA product. The only way I know to do it is to actually create it at the time you want it to run. The 24 hour clock starts then, so it will run at that same time every day. Not ideal if you want it to run at 2 AM granted...
-
Saw this one out there also - maybe you can try it or modify to suit your needs? Custom Report - Top SQL for Database User
-
The short answer is yes and yes. Actually, when you look at DPA, the cpu/memory activity represents a lack of wait within the database engine (the query is executing). The reason we include it in the wait analysis is that we're taking an end-user perspective. They don't care what's happening under the hood, they only care…
-
If troubleshooting steps that GCP laid out don't work out, submit a support ticket. We'll get it sorted out or at least determine why you can't connect.
-
I think you will need to create a custom alert. You can use this one as a starting point: Alert Log Error Alert Then modify the SQL to be something like: select substr(MESSAGE_TEXT, 1, 300) message_text, count(*) cnt from X$DBGALERTEXT where (MESSAGE_TEXT like '%ORA-%' or upper(MESSAGE_TEXT) like '%ERROR%' and MESSAGE_TEXT…
-
DPA currently doesn't do that, but it's a good suggestion. I'll make sure the PM team sees this. You can also post a feature request: Database Performance Analyzer Feature Requests That way, others can vote it up also - helps get visibility...
-
DPA 12.0 looks to be a very stable version. * Upgrade adoption rates higher than any previous version * Support cases and engineering customer issues very low * No known unaddressed bugs that I'm aware of concerning stability The one feedback I've heard is that new navigation was introduced with QPA (Query Performance…
-
By default, that alert will look back at the last 10 minutes for the SQL in question. If the avg wait time per execution is greater than X seconds (as defined in thresholds), it will alert you. You can't go less than 10 minutes as we only grab stats every 10 minutes (so we won't know execution counts to any more…
-
Yeah, I get what you are talking about. The primary reason is that you are not looking at the same dimension. When you clicked on that specific database, it auto flips you back to top SQL Statements for that database (notice the database you clicked on should have gotten added to your breadcrumbs). To see the waits broken…
-
The short answer is no. You can create alerts for resources in DPA 9.5, but storage I/O is not technically a part of resources as such. Feature request?
-
We're getting this data from the vCenter API. We poll every 2 minutes, but get 6 data points within each poll (thus the 20 second interval you see). If this is a false positive, it's coming from the VMware side of the house. I would love to hear if others are having the same issue though...
-
1 DPA installation (server) = 1 repo. No way I know of to map it differently currently. If you want all in one UI, it needs to all be in one repo.
-
You might have to create a custom alert using something like this: Custom Alert - Jobs running longer than usual That way, you can edit the where clause to exclude stuff...
-
Hey Mike, here are some of my thoughts and things for you to look at. Commits is being called out as it's own activity/statement likely because that's how it's being used by the application. Think in terms like the following transaction: begin transaction update table foo set bar = 10 insert into table foo2 values (...)…
-
Haven't had 2005 to monitor in a while, but one thing I do know is that these metrics used to be pulled through WMI using SP_OACREATE procs and scripts. These were disabled in later versions of DPA because in later versions of SQL Server, these metrics became exposed in DMVs which was much more performant. Success Center…
-
8.3 supports SQL 2005. We go about everything we do in a very efficient way. 8 )
-
There is an alert that can send you notification based on a search string. Go into alerts, manage alerts tab, Administrative category, in dropdown, look for SQL Server Error Log Alert. You can select any instances you want plus the search string you want to be notified on. Fairly self-explanatory setting up the alert, but…
-
What kind of benchmark did you have in mind? There are some pretty cool reports you can run to see historical performance. Here's one for SPROCs: DPA Report on Stored Proc Impacts However, we're discussing better ways to aggregate the data. Would you be open for a call so we can get a better idea of what a good benchmark…
-
I'd start by checking whether you have autocommit enabled or not. If not, then your code is opening a transaction (which is a separate statement as seen by MySQL), then when finished, issuing a commit. Keep in mind that autocommit will be disabled if using a start transaction statement even if enabled at the system level.…
-
I'll make an assumption that this is SQL Server, please correct me if not right. We have a couple things going on here. I trust the overall impact (meaning total wait time you are observing from DPA). What might be getting messed up are the stats. There are many reasons why stats for a query will get reset (think…
-
Very Timely. Couple of options currently: Custom Report - Top SQL by Logical I/O per Execution or: Custom Report - Top SQL by Any Metric The top SQL by any metric should do nicely for your request. Or tweak it for your own use. You can bracket the timeframes where PLE is low.
