Here are the respawn times along with the first video’s timestamps:
Since your death lined up with a respawn it’s hard to tell when exactly you died. If you have another recording in which you died we could probably know for sure if dying “fixes” your character.
Yeah. It’s hard to tell for sure! It’s been a theory we’ve been thinking about for a bit, so we’ll have to see and we’re planning on getting some more data on that. Thanks for those specific times!
Sounds more or like what happened to me. My game froze for a good 10 seconds during a Mourningdale invasion yesterday, in Dry Tree.
Thank you. This helps seeing the memory, cpu, etc. during the freeze/pause.
Thanks Jake. Neo is helping me so we can figure out what this is…
Might be helpful if someone with high end-system like Lila Mariposa captures ETW’s as well.
Yup! I’m working on getting her setup with it as well and I will keep you and Neo updated! 
I’m confused, I thought Wars & Invasions would be getting their own instances… Is that no longer a plan of action?
Sadly, this does not fully prevent it (and we have some other theories above, but nothing concrete.)
We’ve recently had a lot of cases where people even this far on the stairs (red arrow) are still crashing with the rest of the gate. It is possible they saw the corruption spawn animation, but it’s not really possible to have everyone play even further up on the stairs. (and most likely something to do with the server side)
I’m hopeful with the new information they can figure out an actual solution
Not really related to this thread, but I see you made a post.
They will.
@JakeL I think it may actually be relevant. The lag, game crashes in Invasions has been going on for a considerably long time. The convention of thought on this was that if your server had wars simultaneously, then this was the root cause. however we’ve seen these incidents still occur without concurrent wars in progress. So it seems logical that creating “instances” for Invasions in addition to wars would or should be the answers. It seemed as though AGS was working in this direction. Unfortunately it appears as though these efforts have stalled. To now see Dev’s asking the community to help them solve this long standing lag/crash issue, makes me wonder if they’ve given up. The “Instances” certainly seemed like “The” answer.
I’ve had this crash on servers with varying levels of performance. Crashes at max capacity & crashes at half capacity. Creating instances for these events will certainly help overall performance. It could help, but it’s unlikely to solve this issue. Who knows, perhaps a future version they’re actively working on doesn’t have this problem.
I think what the developers main goal was to detach the wars from the open world so that it doesn’t see a performance drop during these events. It should help isolate performance issues with wars and invasions themselves which this seems to be one of them.
I just wanted to mention that this has been happening on El Dorado for about the past month. It’s usually a pretty big chunk of people and they will either crash fully (game closes and have to reopen) or will freeze and come back like some of the videos above.
This continues to plague invasions. I am on US West; El Dorado and have had the same situation on multiple invasions.
Last night (09/05) it was the back gate that had multiple people freeze (more than 5) and either drop to desktop or force-end-task as they were stuck. I attended two invasions last night and can’t remember which one it was, but I think it was the 8PM Brightwood Invasion.
Tonight (09/06) it was the mid-gate in front. More than 5 players at the same point in the invasion. 7PM Restless Shore Invasion.
We have also noted that if you are in a group at the time of the Invasion invite, if you accept the invite, it will place you in a group… then automatically kick you right back out. This has happened several time. We now prep for invasions by making sure we leave groups a few minutes before invites can go out.
Yep I was shot calling mid tonight on Eld and my wife and I both hard froze and had to task mngr force close. Then I froze another 2 times but didn’t have to force close.
This is a separate bug, not related to this thread. I would recommend making another thread with an actual bug report for this, as I don’t believe this has actually been reported correctly yet.
If you continue to crash, I’d heavily recommend getting the Windows Performance Recorder installed as well as recording software so that you can get the specific debugging files they are looking for. Saying it’s crashing doesn’t really helps anyone, it’s happening to all of us.
I would recommend the same thing. If you can record the crashes and freezes on something like OBS as well as Windows Performance Recorder, it sounds like it would be very helpful to the devs. My server has been doing so as often as we can with Neo.
Hi Jake, thank you (and all) for helping on this, it has helped tremendously!!! Can you confirm that all of the ETL captures provided were from Invasion?
At this time we think we have enough ETL captures for Invasion specific issues.
“We have also noted that if you are in a group at the time of the Invasion invite, if you accept the invite, it will place you in a group… then automatically kick you right back out. This has happened several time. We now prep for invasions by making sure we leave groups a few minutes before invites can go out.”
Thank you @Drumana, I am asking if we are aware of this issue and if not will make sure it gets to the right team.
Hi @Prasinos333, War and Invasion instancing was turned back on which will help performance in general. The specific freezing/client stall issues discussed here appear to be client issues which we are actively looking into.
Yes! All of the ETL captures have been during invasions.
Great! Let us know if you need anything else. We’re happy to help!
I just wanted to take a moment to thank @JakeL, @Pwent and others who provided us with their ETL captures. Your hard work was instrumental in helping us fix the issue and yesterday’s patch was a direct result of it. We improved memory management for particles to fix these intermittent client hangs that could occur during certain race conditions. We hope this will alleviate what you’ve been seeing but in the event you’re continuing to see the these performance issues, feel free to reach out to us and we’ll work with you similarly to address anything else that might be found in additional captures provided.
