We use cluster active gates(public) for both agent-based and agentless RUM implementations.
Some of our applications are configuration heavy example: 150 session and action properties configured.
We saw errors with 431 response code on the RUM beacon calls, and although Dynatrace support initially thought it was not an error code the ActiveGate returns, but rather it could be something the underlying Active Gate server/OS is responding with.
However, upon further investigation, we concluded that this response code occurs when the request header size on the beacon calls goes above default, which is 8 kb (This info is not available in any of the public-facing documentation).
The fix for this issue was to add a custom property configuration to the active gate to increase the permissible header size. In our case, we increased it to 16kb.
We saw that these errors cannot be identified via any logs or self-monitoring we have on the active gates. Can this monitoring capability be made available so that we can proactively increase limits if needed to avoid RUM data loss for critical apps?
