I know this has nothing to do with web SDK but there are no sub section which fit for this kind of topic so please forgive me to post that here and feel free to move it to a better section.
Description
We are a Japanese company with 100% of our activities here in Japan. So our zoom account is setup to enable JP as available region for gateway / data center ping when connecting to Zoom. US is the default and non modifiable fallback.
From mid August to beginning of September, most if not all of our meeting were hosted in Japan region which was perfect. Then from beginning of September we noticed that most of the meeting started to be hosted in US instead.
From beginning of September, our users started to report audio quality degradation on Zoom services, and we suspect that being redirected to US data center is part of the cause. As you can see in the following screenshot, users can have more than half a second of latency to reach these regions.
Since few days, we changed our settings to enable Hong Kong region as this is closer so should be better. But this is not going to fix the root cause.
Here are our concerns regarding this situation:
Why a 404 is returned on the ping ? I guess your did not remove that data center as we sometime can join it. So this is to say “i’m not available”, right ? Can you instead return a 503 which is the code to say that you are not available due to overload ?
Is it possible to increase resources allocated to Tokyo data center / gateway to accept more traffic for this region ?
Please understand that this has an impact in term of service quality delivered to our customers.
It looks like it’s only during business peak hours that we cannot use Tokyo. Right now at 8:24 am for example, all meetings are hosted on TK data center.
But starting from 10am to probably around 10pm it will be 95% out of Japan.
So the resources required to absorb the load in JP appear to be underestimated compared to the actual demand.
Is it possible to review it ?
There are three account and two are free trial accounts.
Free trial accounts have lower prioritized using our infrastructure due to the fact its free.
Can you provide one of the the meeting id as an example from what you are stating?
If those informations are sensitive then please use developersupport@zoom.us to share us.
Basically this is by design and we will have to keep this as it is for now.
As in a high level explanation, handling WebClient/WebSdk calls will first seek for all available gateways and and tries to choose the least loaded.
If the gateway in that location is full we will reply with 404.
Is it possible to increase resources allocated to Tokyo data center / gateway to accept more traffic for this region ?
We are currently in the faze of enhancement in the performance and capacity perspective and have added your request as well.
This may take a while and for now our recommendation will be, from your admin portal, deselect HK and instead choose Singapore. (you can keep other location as it is)
WebPortal: ADMIN -> Account Settings -> In Meeting (Advanced) -> “Select data center regions for meetings/webinars hosted by your account”: uncheck HK and check Singapore
Thank you for your reply and so sorry for keep you waiting.
Seems that still our APAC resource was not enough to reroute using Singapore as well.
We are discussing this issue internally on how we can resolve.
We will try to get back to you asap.
13 days without any news and Singapore data center is still not available… what’s the hell is going on seriously ?
Since yesterday, almost all our users being hosted on Hong Kong are reporting extremely bad audio and video performances (choppy, lags, freeze), and we can easily figure quite a huge latency from zoom meeting dashboard for those users.
What’s going on with HK data center and Zoom network ?
Another point is, now that you have Osaka (OS) and Tokyo (TY) data centers, we can figure that sometime users go on Tokyo and sometime on Osaka (very low percentage)
But this is only when TY Gateway reply with 200 instead of 404…
We have reviewed the situation with our engineering and also is planned to increase the capacity for Japan location.
Once the deployment is done we will let you know once more.
I guess you’re talking about reducing the schedule to November to increase JP gateway capacity as explained in the email exchange.
If not, could you re-explain what are you planning ?
