This would be expected. We’ve just experienced the same thing using workers to respond to preflight requests at the edge. Our argo bandwidth exploded about 8x what it is without workers. This makes absolutely zero sense. In both cases, payload requests will transit argo. I have no clue how this results in increased bandwidth.
I find it really bizarre that we can’t get an answer to what should be a simple question. We have asked in Discord and have subscribed to a paid support plan just to get an answer to this.
When Argo is enabled, a worker that simply forwards requests to origin and does nothing else results in a large increase of bandwidth, despite the fact that the same number of bytes should (in theory) be transiting Argo. Nobody from CF has been able to explain this yet.