Add an end-to-end test that amplifies the response size

[eddyashton]

We have end-to-end tests that submit extremely large requests to a CCF node (e2e_common_endpoints.py:test_large_messages), in the body and headers and various configurations, and other end-to-end tests that produce extremely large KV writes (batched.js). I think one case that we're missing is in unusually-large, enclave-generated responses. In particular, when a HTTP response approaches ringbuffer size limits, without being previously gated by ledger or request size limits, how does a node respond.