Existing documentation URL(s)
https://developers.cloudflare.com/cloudflare-for-platforms/workers-for-platforms/configuration/outbound-workers/ says:
Outbound Workers sit between your customer's Workers and the public Internet. They give you visibility into all outgoing fetch() requests from user Workers.
Log all subrequests to identify malicious domains or usage patterns.
When an Outbound Worker is enabled, your customer's Worker will no longer be able to use the connect() API to create outbound TCP Sockets. This is to ensure all outbound communication goes through the Outbound Worker's fetch method.
The page repeatedly asserts that outbound workers is usable as a security mechanism, but then at the very ends mentions:
Outbound Workers do not intercept fetch requests made from Durable Objects or mTLS certificate bindings.
This seems like a significant caveat to "all"? Couldn't a malicious worker simply arrange for their outbound fetches to be sent by a Durable Object to avoid being subject to any observation?
What changes are you suggesting?
Elaborate on how to "Log all subrequests to identify malicious domains or usage patterns." when customers are using Durable Objects, which reportedly bypass outbound workers.
Additional information
No response
Existing documentation URL(s)
https://developers.cloudflare.com/cloudflare-for-platforms/workers-for-platforms/configuration/outbound-workers/ says:
The page repeatedly asserts that outbound workers is usable as a security mechanism, but then at the very ends mentions:
This seems like a significant caveat to "all"? Couldn't a malicious worker simply arrange for their outbound fetches to be sent by a Durable Object to avoid being subject to any observation?
What changes are you suggesting?
Elaborate on how to "Log all subrequests to identify malicious domains or usage patterns." when customers are using Durable Objects, which reportedly bypass outbound workers.
Additional information
No response