Slowdown of interface

jonstahura

I have 8 servers on my sql monitor.

The response time has slowed down SO much.

When I try to open alerts from the inbox, it takes 30 or more seconds to open.

Is there a reason for this? Can I speed this up?

mipe10

I have the same problem! I have registered a case to get some help.

Brian Donahue

I've sent some debugging steps - basically this is not the kind of thing we can answer without detailed information about the base monitor and website processes.

If you have lots of users connecting to the website (more than 10) you may want to upscale to IIS if you're using the built-in web server.

jonstahura

well it is just me.... so i don't see users being an issue.

I was doing some testing.. it seems that if I remote into the monitor server and open a browser to localhost... it is WAY WAY faster...

Everything else seems to run ok... just the alerts take FOREVER..

mipe10

Hi,
I can confirm what jonstahura is saying about the local login and execution on the actual machine containing the installation. It's lightning fast compared to remote connection.

Brian Donahue

Seems to implicate your network connection. None of RedGate's code is going to care if the requester is local or remote.

I'm aware also that the Javascript is quicker on some browsers than others.

mipe10

The problem was chrome! Changes to explorer(64 bit) and the problem went away.

jonstahura

i will test it with IE to see if i have the same results..

i also have been using chrome... i hate using IE.. this is the 3rd application in my company that "FORCES" me to use IE.. grrr

although the commercial song for IE is catchy... lol

jonstahura

i hate to say it.. but ti works...

something must be happening in chrome... .. so depressing... IE is nice and snappy compaired.. i have not had ANY of the wait time that hasbuilt up in chrome.

Chris Spencer

I'm not seeing any issues with Chrome internally, monitoring 10 servers. We fully support this browser and in general get far less issues with it than we do with IE. I'll keep an closer eye on Chrome to see if I can reproduce this issue.

If you have an aversion to IE then SQL Monitor also works with Firefox, Safari and Opera.

Regards
Chris

jonstahura

i can consistently recreate the issue. If you would like to do a remote or something to see if you can track something down, i would be willing to be a test bunny.

It really creates a wait when opening up alerts.

henkmeulekamp

And on your advice tested with IE9 (win7) and its fast.

Definitively an issue with chrome and Sql Monitor web app.

It is especially when hitting the alerts tab and opening an alert.

We only have 3 monitored servers, sql monitor on a complete separate infra.

Chris Spencer

A quick update to say that I'm now seeing this issue in Chrome and have raised a bug (ref: SRP-7768).

We'll look into this and hopefully improve performance in a future release. In the meantime IE9 and Firefox are definitely a better choice for Alert Inbox work.

Many thanks
Chris

Baguck

Thanks for the thread and bug - I am noticing similar results.

Freman

Firefox is much more fast than IE to SQL Monitor

Chris Spencer

Hello

Is anybody still seeing performance issues with the Alert Inbox when using Chrome?

We're currently investigated this but have been unable to reproduce in recent versions of Chrome. So we're thinking that a Chrome update might have fixed it.

Regards
Chris

PDinCA

Upgraded to 3.2.1.231 on Feb 19th. Initially thought it was a horrible rev because:

1) Performance Data is forever "loading..."
2) Cannot use the Older/Newer links
3) Opened a new tab and when closed it left its imprint over the Monitor page, or a grey page background.

I pulled that bug report - until today, when I'm back to those issues...

Am certainly on the latest Chrome Version 24.0.1312.57 m.

Series of actions today:

Monitor was open at the Overview page for the cluster.
. Configuration ==> Licenses
. Alerts
. Detail a Job Failed alert. Cleared it.
. Detail a R: Drive space alert. Configured the Alert to disable it.
. Detail a Q: Drive space alert. Configured the Alert to disable it.
. Detailed a "HarddiskVolume" Alert, which is VERY confusing as all it says is 0 used, 0 free, etc. Don't even know what it is because the Performance Data is still "Loading...".

The page is STONE COLD DEAD apart from the title bar's "Help" , which at least allowed me to get to the forum. Yesterday I could click the main Overview/Alerts/... tabs and change the page, but the content was nixed after the first "Loading..." issue.

I just tried to view the Monitor page - my Forum page is what is showing, although the mouse pointer does change as I move over the page's controls, so something is registering despite being unusable.

Perhaps what was done to accommodate IE10 (IE% = Browser of last resort) messed up Chrome... I was on the immediately prior rev of SQL Monitor and had no rendering or freezing pages issues whatsoever.

Underlying OS is Windows 2008 SP2.

Any other assistance I can give?

If this isn't resolved or in analysis, I'll have to revert...

Chris Spencer

Hi Stephen

I've investigated this and have seen some performance issues with the Disk Space alert where the Performance Data section doesn't load and the page eventually times out. I only see this when using Chrome or IE9. Firefox seems to be far quicker at this particular task.

I'm not seeing points 2) or 3) however. Maybe my system isn't struggling quite as much as yours. The system that I'm seeing this on has a 200GB data repository by the way, my other systems with lower purge settings are fine. It would be interesting to know the size of your SQL Monitor repository for comparison purposes.

Anyway I've raised a bug for this (ref: SRP-8541) and we'll investigate further.

Regards
Chris

PDinCA

Hi Chris.

Thanks for the swift follow-up.

My repository is only 20GB. I have a 14-day max retention and some at 7 days. I only monitor one SQL Cluster (Budget-distressed for any more - need 3 more!).

The unusable page is probably due to the performance data not loading. The page-litter is also reproducible under the same "Loading..." condition, so they're symptoms, not separate bugs methinks.

Regards.