Hello Robert Heynen, the timeout attribute of the forward request policy documentation mentions that values greater than 240 seconds may not be honored, because the underlying network infrastructure can drop idle connections after this time. So, try setting a shorter timeout to see if the policy works as expected in both the scenarios. Also, in the Consumption tier, the timeout settings might be more flexible or have different default behaviors compared to the Basic tier. This might emanate from the fact that the Consumption tier is mainly designed for serverless and event-driven workloads, which might inherently support longer timeouts better than the Basic tier. You can try putting in a support ticket with Microsoft if the issue persists.
Please do not forget to "Accept the answer” and “up-vote” wherever the information provided helps you, this can be beneficial to other community members.