Error handling for aws bedrock and openai response

[yuzisun]

• Add logic to handle error response from upstream aws bedrock or openai, currently it is assuming the response is 200 and fails on response body unmarshalling.
• Add tests for error responses from upstream

time=2025-01-19T18:34:38.104Z level=INFO msg="starting external processor" version=8056bb8bcfe484ca6f6007b021ea69448b3073f4
time=2025-01-19T18:34:38.104Z level=INFO msg="loading a new config" path=/etc/ai-gateway/extproc/extproc-config.yaml
time=2025-01-19T18:34:38.104Z level=INFO msg="start watching the config file" path=/etc/ai-gateway/extproc/extproc-config.yaml interval=5s
time=2025-01-19T18:34:42.422Z level=DEBUG msg="request headers processing" request_headers="headers:{key:\":authority\"  raw_value:\"localhost:8080\"}  headers:{key:\":path\"  raw_value:\"/v1/chat/completions\"}  headers:{key:\":method\"  raw_value:\"POST\"}  headers:{key:\":scheme\"  raw_value:\"http\"}  headers:{key:\"user-agent\"  raw_value:\"curl/8.7.1\"}  headers:{key:\"accept\"  raw_value:\"*/*\"}  headers:{key:\"content-type\"  raw_value:\"application/json\"}  headers:{key:\"x-ai-gateway-llm-backend\"  raw_value:\"kserve-llama-backend\"}  headers:{key:\"content-length\"  raw_value:\"113\"}  headers:{key:\"x-forwarded-for\"  raw_value:\"10.244.1.40\"}  headers:{key:\"x-forwarded-proto\"  raw_value:\"http\"}  headers:{key:\"x-envoy-external-address\"  raw_value:\"127.0.0.1\"}  headers:{key:\"x-request-id\"  raw_value:\"41605e8c-7ef8-4fbc-b1f9-a8f01740746e\"}"
time=2025-01-19T18:34:42.422Z level=DEBUG msg="request headers processed" response=request_headers:{}
time=2025-01-19T18:34:42.422Z level=DEBUG msg="request body processing" request="request_body:{body:\"{\\\"model\\\": \\\"dsp.llama-3.3-70b-instruct\\\", \\\"prompt\\\": \\\"Write a poem about colors\\\", \\\"stream\\\":false, \\\"max_tokens\\\": 100}\"  end_of_stream:true}  metadata_context:{}"
time=2025-01-19T18:34:42.422Z level=INFO msg="Processing request" path=/v1/chat/completions model=dsp.llama-3.3-70b-instruct
time=2025-01-19T18:34:42.422Z level=INFO msg="Selected backend" backend=kserve-llama-backend.ai-gateway
time=2025-01-19T18:34:42.422Z level=DEBUG msg="request body processed" response="request_body:{response:{header_mutation:{set_headers:{header:{key:\"x-envoy-ai-gateway-model\"  raw_value:\"dsp.llama-3.3-70b-instruct\"}}  set_headers:{header:{key:\"x-envoy-ai-gateway-selected-backend\"  raw_value:\"kserve-llama-backend.ai-gateway\"}}}  clear_route_cache:true}}"
time=2025-01-19T18:34:57.425Z level=DEBUG msg="response headers processing" response_headers="headers:{key:\":status\"  raw_value:\"504\"}  headers:{key:\"content-length\"  raw_value:\"24\"}  headers:{key:\"content-type\"  raw_value:\"text/plain\"}"
time=2025-01-19T18:34:57.425Z level=ERROR msg="upstream returned non-200 response"
time=2025-01-19T18:34:57.425Z level=DEBUG msg="response headers processed" response=response_headers:{}
time=2025-01-19T18:34:57.426Z level=DEBUG msg="response body processing" request="response_body:{body:\"upstream request timeout\"  end_of_stream:true}  metadata_context:{}"
time=2025-01-19T18:34:57.426Z level=DEBUG msg="response body processed" response=<nil>
time=2025-01-19T18:34:57.426Z level=ERROR msg="cannot process request" error="cannot process response body: failed to transform response: failed to unmarshal body: invalid character 'u' looking for beginning of value"