Hey there, we’re on 2.02 and use lots of short duration snapshots in events to manage tight ducking as opposed to side-chaining, as it allows for a lot of control and is an easily monitored mix behavior when profiling.
Unfortunately, the profiler hangs on to snapshots which have recently been used for around 10 sec, even after the event which invoked them has died. This makes the bottom right hand side of the profiler that displays current snapshots difficult to read as it fills up with a ton of dead entries.
Is there any setting to have the profiler visually flush these out more quickly?
Unfortunately, as of the time of writing (July of 2024), there is no setting to change how long snapshot instances remain visible in the profiler window after being stopped; they always remain visible for 10 seconds after stopping.
Thanks Joseph! I’d like to request this as an option for the authoring client at some point down the line, I think it would really increase clarity during profiling. Is there any engine-bound reason why they would need to stay visible in the profiler for that long? Are they utilizing any resources during those 10s?
No worries! I’ve already added the idea to our feature/improvement tracker and added you to the list of people interested in it.
No, not at all. Snapshot instances consume almost no resources to begin with, and stop consuming even that small amount immediately upon ending.
The profiler displays snapshot instances for ten seconds after they end purely because it makes it easier to tell that a snapshot instance was created, even if the instance only lasted a very short time. We consider this an important feature, as it allows users to more easily identify where their game code is mistakenly stopping snapshot instances very soon after creating them.
To put it another way, “blink and you’ll miss it” is not a phrase you should need when using a diagnostic tool. Still, there should be no problem in allowing you to customize the amount of time a snapshot instance remains visible after it ends.