Hi everyone,
We’re seeing a strange issue when a large number of users join our webinar using Meeting SDK for Web 6.2.0 (Client View).
All users are on mobile, using our React web app inside a React Native WebView:
- iOS: WKWebView
- Android: Android System WebView
@zoom/meetingsdk: 6.2.0- Webinar attendees
- Zoom SDK assets loaded from
source.zoom.us
For affected users, the SDK logs show:
success load jsmedia
pre load wasm success: https://source.zoom.us/6.2.0/lib/av/video.simd.wasm
pre load wasm success: https://source.zoom.us/6.2.0/lib/av/audio.simd.wasm
success load webim
success load tp
POST https://zoom.us/api/v1/wc/info status_code: 0
You have been limit by zoom, verfy recaptcha and try again, open https://zoom.us
{ errorCode: 1, errorMessage: "Fail to join the meeting.", method: "join",
result: null, status: false }
The SDK assets load successfully, but /api/v1/wc/info returns status_code: 0 and the join fails with the Zoom limit/reCAPTCHA message.
We’ve ruled out a few things:
- We were well below our licensed attendee capacity.
- The same signature works for users who successfully join.
- It is not specific to one device/browser.
- Some users who failed initially could join successfully a few minutes later.
Our main suspicion is rate limiting, possibly based on public IP. Most users are on mobile networks, so many devices may share the same public IP through carrier NAT.
A few questions:
- Has anyone experienced this when using Meeting SDK at a large scale?
- Is there a documented rate limit for
/api/v1/wc/info? Is it per IP, account, SDK app, or something else? - Is there a way for the SDK to actually display the reCAPTCHA so the user can complete it and continue?
- Zoom Support suggested enabling “Require solving a CAPTCHA for guest users.” Has anyone tried this with Meeting SDK/WebView? Did it help?
- Does the “Show a Join from your browser link → On mobile browser” setting affect the Meeting SDK join flow?
Any information about the actual limit or recommended approach would be really helpful. We want to avoid simply increasing delays without knowing what limit we are trying to work around.
Thanks!