Should be resolved by now - a leftover of the old cluster apparently was connected to the boost frontend endpoint still.

3 Likes