check_connection_for_update

takes too much time and ends up with a 502 error (probably a timeout) on the Facebook Marketing v3.4.8 connector; is there a workaround ?

The knowledge sources don't contain specific information about version 3.4.8 of the Facebook Marketing connector or a dedicated fix for

check_connection_for_update

timeouts in that version. However, based on related issues documented, here's what's known and what you can try: Known Issue This is a recurring problem with the Facebook Marketing connector. A similar timeout issue was reported in v3.3.29, where

check_connection

(schema refresh and new connection setup) timed out with

HTTP 504

/

Gateway Time-out

errors. [GitHub Issue] Potential Workarounds 1. Check if it's an infrastructure issue The reporter of the v3.3.29 issue found that running locally worked fine, suggesting the problem may be related to the Kubernetes/infrastructure deployment rather than the connector itself. If you're on a self-hosted K8s setup, investigate: • Ingress timeout settings (nginx, etc.) • Temporal activity timeout configuration [GitHub Issue] 2. Increase the timeout If you control your ingress or proxy layer, increase the timeout beyond the default. The original reporter increased it from 1 to 20 minutes, though that alone didn't resolve their issue. [GitHub Issue] If you're behind nginx, note that a shared nginx ingress with a hard 60s timeout (that cannot be increased) is a known blocker. [GitHub Issue] 3. Reduce the scope of data being checked For Salesforce (and likely other connectors), filtering to only the tables/streams you need can reduce check connection time. The same principle may apply to Facebook Marketing — limiting the ad accounts or streams configured in the source. [Forum Post] 4. Retry until it completes within the timeout window A suggested workaround for gateway timeout scenarios is to keep retrying the check until it completes within the timeout window. [GitHub Issue] Root Cause Context The Facebook Marketing connector uses async jobs for Insights, which can sometimes get stuck for hours on Facebook's side. The connector has retry/split logic (splitting by campaign → adset → ad level), but this can make the overall check take significantly longer when Facebook's API is slow. [README] I'd recommend opening a GitHub issue specifically for v3.4.8 if you haven't already, as the knowledge sources don't confirm whether this is a known bug in that specific version or if a fix has been released. 👍 0 button 👎 0 button ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.