Region/Estate tools "Top Scripts" vs the Stats Window

[Innula Zenovka]

I've just noticed that the figures for the number of scripts and the script time on my sim differ considerably between the stats window and the "top scripts" shown in the Region/Estate tools.    The figures in the stats window are about three to four times those given by the Region/Estate tools.    This is the case in both Catznip (which is based on 2.8) and the latest beta viewer.    

I don't remember there being such a wild discrepancy in the past.   Are they now measuring different things, or is something bust?

[Rolig Loon]

Top Scripts is a snapshot, so it may get lucky and catch a script when it's actually doing something or (more likely) while it's idling between triggered events.  The stats bar is a dynamic display, with some time averaging too, I suspect.  As a result, it's more likely to see more of the momentary peaks when active scripts are actually doing something.  At least, that's how I have rationalized it -- as a sampling issue.

[Innula Zenovka]

That's changed, though, Rolig.   According to the wiki

TOP SCRIPTS: Prior to server version 11.06.17.233176 (released 2011-06-28) the times reported represent only a brief instance in time and are inaccurate. Server 11.06.17.233176 changed it so that the times reported are now the average script time per frame for the last 30 minutes for all scripts on the object. If the object has been in the region for less than 30 minutes the number reported will be the average time per frame since the object entered the region.

I'm not certain, and would appreciate confirmation or correction on this, but I think, from the wiki articles, the Stats Bar tells me how many scripts are running and Top Scripts tells me how many scripted objects there are (though I've never noticed so great a discrepancy before).   But this doesn't explain -- at least not to me -- why what's supposedly "the average script time per frame for the last 30 minutes for all scripts" in the region is about a third of the figure in the stats bar.

[Void Singer]

assuming they are bothe averaging (LL has gone back and forth on making top scripts average or snapshot), it could be that they are averaging over different time frames.... for instance if top scripts averages for 30 mins, and stats for 2, if script time shot up in the last 3 mins, top scripts would only register 1/10th the change. so if the script time had quadrupled (such as a monster scripted av TP'd in), it'd be possible to see a huge difference between the two.

[Innula Zenovka]

Thanks, Void.   I guess I will need to keep an eye on this and see what happens.   I'd be very grateful, in the meantime, if other people with access to Top Scripts could take a look and say if they see any big discrepancy between the figures there and the stats bar, particularly on the main server channel.

There was certainly something very odd happening on the sim in question yesterday; time dilation (as reported by the stats bar and as was obvious from trying to do anything) started going all over the place and, when I checked, an apparently innocuous scripted tree that's sat there for a couple of years hardly taking up any script time seemed suddenly to have gone berserk for no apparent reason and was somehow showing a time of 60+.  I disabled scripts in it, but that didn't do much good, so I restarted the sim, after which things stablised but I first noticed this big discrepancy, which was still there a couple of hours after the restart. 

All the other figures in the Top Scripts window looked pretty much as I would have expected (though the total figure was a lot lower than it has normally been) and the stats bar looked pretty much as usual, too -- it's just there was now a big discrepancy.    I've never really paid much attention to Top Scripts until now, unless there's a problem or I'm looking for something I've lost, but now I'm wondering.

[Six Igaly]

Top Scripts 1.5 ms

Stats  4.7 ms

Second Life Server 11.10.18.243270

[Innula Zenovka]

Thanks.  So, again, one's about a third of the other.   So it looks as if they may now be measuring different things (or reporting the same measurement in different ways?).