Definitely, it is a big issue that needs to be fixed as it is also impacting the growth of Decentraland.

Does Decentraland need to allocate more resources to ensuring scenes can handle higher capacity of users? This proposal is now in status: REJECTED. Voting Results: * Yes 49% 1,286,139 VP (67 votes) * No 0% 0 VP (0 votes) * Invalid question/options 51% 1,288,049 VP (25 votes)

I agree this should be a priority. Not only have scenes been crashing with minimal users and taking a long time to load when you walk up to them, but also just loading Decentraland and getting in-world seems to take a lot longer the last couple of weeks.

The game is way too laggy. Scenes needs to be improved... Can't go 5 parcels away before, the scenes have a hard time to load... We need to improve the game drasticly if we want DCL to suceed. 100% we need to have focus on this, and soon! It's always hard, when onboarding new users and land-buyers, land-renters for me.. It's always a discussion when they take a look around, how tought it even is to walk around. So i vote yes, we need dev's focus on this. For a better DCL.

I am not technical at all however I do go to A LOT of the parties and the crashing and fatal errors have become very frequent. I attempted to attend the Jagermeister Lesbian Bar Project and crashed 5 times in about a 20-minute period. Yesterday I had to refresh and jump back in during the Rage event and had to completely reboot my computer at Metazoo multiple times because everything just froze. During the opening of The Holy Ones Temple, I documented 19 times that I had to refresh to get back into the event. Almost every time that I visit Butterfly Prawn I crash at least once. Even in Petaverse when there are only 1 or 2 users I have had to refresh. It seems to be a DCL wide problem, but I wonder if the influx of new emotes have caused a much more frequent issue. I am really happy we are having this discussion; I just hope this will lead to a real proposal with a valid solution in the near future.

[quote="Unknower, post:39, topic:15826"] Have you actually surveyed *every* scene operator or been to *every* scene and seen this happen *every* time? Disproving that hypothesis would actually be a pretty valuable test to do. If you can find scenes where this *isn’t* happening, that could provide some clues. Has anyone had this issue at any of Samsung’s recent events? They have a large build with multiple video screens and I haven’t personally had or heard of any issues with crashing there. [/quote] Sounds pretty impossible when you cant find 80 people in Decentraland. But you want us to walk scene to scene together to find the issue? Ofcourse the scenes with no dev will probably not encounter any issues. But the question is, why are these developed scenes so easy to "break?"

Last couple of days its just unplayable to be in DCL, the realms cant keep more than 40 people to not crush i see ghosts all the time even im a poker player i have a issue i cant see tables and last few days i need like 10 to 20 minutes to load a scene on poker, its just super unplayable to be and play in this metaverse.

[quote="web3nit, post:40, topic:15826"] This is issue that need to be solved asap, Decentraland must handle users without lagging,bugging,crashing. While DCL gives grants for games, people cant play normally that games. [/quote] I agree. [quote="web3nit, post:40, topic:15826"] Lol remembered my first proposal [/quote] Is there a better process to go with then? I am sincerely asking, because this is ridiculous, that the foundation wants to point fingers at the Users and scene devs, or too many triangles lol. How about they upgrade their system to make it so it can't "break" as easy.

Just to add some anecdotal evidence, I attended three events yesterday and experienced similar issues at all three. The first event (Rage) I arrived after there were over 20 people already there, and had to refresh multiple times to load just the live stream music and have my avatar participate in autodance. At no point did more than 5 other avatars load in, and I went no more than 5-7 minutes without getting the fatal error screen. After giving up there I attempted to go to Metazoo, which had more attendees than Rage at the time, and loaded in unable to move my avatar, with no UI, no music, and invisible avatars. On the third refresh, the scene elements worked but no one else would load, and I got the fatal error less than a minute later. Then, I hopped over very early to the Holy Ones opening event that at the time only had 8 users and arrived with no issues. I stayed until the event began and didn't crash or experience any issues for over 45 minutes until the crowd started growing and then the crashes started again. We experienced similar issues at a Dice Masters event on Sept. 30th, and weren't using live streaming so I don't know that it can be explicitly connected to live stream events.

While I agree this proposal solves nothing, only to bring awareness of the issue, there is a need for someone to track down the problem and provide a solution. If its a DCL server issue then provide a work around or fix, if it turns out that it is a scene builder problem then provide guidance to scene builders so they can adhere to guidelines and not have this crashing issue. This seems to be a recent development, not limited to just one scene. I have been frequenting Metazoo Friday Night Club Night for over a year, but it has been crashing like this for only the past month or so. Yesterday's Holy Ones party and the Frost Monarch burn party on Sept. 30th at Dice Masters are other examples of this crash. It is weird that these crashes do not occur for me when I use a browser VPN plugin to "disappear". When using this VPN I cannot see anyone else nor can they see me, I am totally alone in scene. This is the only way I can attend parties which sucks because I cannot communicate with other attendees. I can go to the most crowded party realm in DCL with my VPN on and no crashes, but soon as I turn it off and people start loading in for me, that is when the scene crashes. I am by no means a technical person with this kind of issue, but IMO that tells me it is an attendance issue, not a scene builder issue. It is quite strange that this issue seems to have started around the time emotes were allowed, I would suggest a more technically inclined person investigate this.

Lol remembered my first proposal [https://governance.decentraland.org/proposal/?id=9f2959d0-637a-11ec-8188-4352ce3d30e7](https://) :joy: My new macbook pro handles with DCL pretty good but before that on pro 2017 it hardly crashed and got heated asap. This is issue that need to be solved asap, Decentraland must handle users without lagging,bugging,crashing. While DCL gives grants for games, people cant play normally that games. Also I saw some FUD-like news [https://www.coindesk.com/web3/2022/10/07/its-lonely-in-the-metaverse-decentralands-38-daily-active-users-in-a-13b-ecosystem/.](https://) Maybe we can have law suit on trash talkers ??

Have you actually surveyed *every* scene operator or been to *every* scene and seen this happen *every* time? Disproving that hypothesis would actually be a pretty valuable test to do. If you can find scenes where this *isn't* happening, that could provide some clues. Has anyone had this issue at any of Samsung's recent events? They have a large build with multiple video screens and I haven't personally had or heard of any issues with crashing there. This issue has a much better chance of being taken seriously, found, and fixed if we provide detailed reproduction steps like @menduz asked for, rather than speaking in hyperbole and absolutes like "every", "always", "only", "never", etc.

But why is every scene having issues?

They keep acting like it's not a problem. There were 3-4 new users at the Holy Temple launch which basically quit DCL. They say it's the scene devs or many other factors. But why not enchanted DCL so these factors are less of issues..

This was also a problem at the Jagermeister Lesbian Bar event yesterday. People kept crashing repeatedly. A lot of new users were there, and it was frustrating to try and tell them this is just "business as usual". I crashed around 10 times. Some people have found that certain emotes will crash another avatar if the user does it enough. Is there any possibility that the sudden influx of emotes are having an effect on crashing avatars?

They are using LivePeer, you can configure LivePeer to transcode to a single resolution and bitrate, that will likely solve the issue, David might try the next event on Monday with my server to confirm the crashes are due to the streaming through LivePeer

Actually, they are decentralized in that you can just set up your own infrastructure and make it as robust as you want. Then propose to add it to the catalyst network. https://docs.decentraland.org/contributor/tutorials/how-to-run-a-catalyst/

Is this happening when there are multiple video screens? If so, are all instances of the video screen using the same Material instance and VideoTexture instance? If an error is being thrown inside the browser console, this has nothing to do with DCL's server resources and points to something potentially being wrong within the scene code. We can't allocate DAO funds to solve an issue unless we completely rule that out and actually can point to what problem would need to be fixed. If we are seeing [5xx HTTP responses](https://en.wikipedia.org/wiki/List_of_HTTP_status_codes#5xx_server_errors) from Decentraland APIs, that would be evidence of a server issue.

I do believe we were streaming on rtmp server for the Holy Temple. Please check with David. However, if this is common, then how can Decentraland embrace this behavior? I am stuck on this - if we are blaming many scene devs putting their scene together improperly, is there a commonality?

I'm pretty sure the problem is related to the streaming solution used. They are likely transcoding into multiple resolutions and the client doesn't like it (never has). That problem was not an issue in the past because most people were using services that were not transcoding, means the streams were in a single resolution and bitrate (like my service or `rtmp-server.com`), but with LivePeer being easy to use and accessible, several people started using it and LivePeer is transcoding the flux it receives into several resolution and bitrate, the client is switching between them depending on your connection, which in theory is good, but was never properly supported by the client.

[quote="mattimus, post:17, topic:15826"] When using 720p with 2500kbps it was much more stable. [/quote] This is the quality Rage uses as well as the Holy Temple was using. Both were crashing!

This is a poll to move forward. This is the process they have for the foundation to make changes. This is a foundational change. Otherwise I would.

We had a few new users coming to Decentraland for the first time, heres one of them from tonight  Then we wonder why we cant get new users on board...

Haha I hear ya bro but again the best way to find solutions is to figure it out or team up with someone that can. Because if not then we are wasting everyones time and using this DOA process as a place to complain and vent which I don't belive it is intended for. Again intention is good and I look forward to voting on your solution.

Thank you!!! Much appreciated!!