MCPX MCP Gateway

@legacy Feature: Lunar Proxy MetricsCollector Diagnosis Background: Starts the Proxy Given API Provider is up # The next 2 steps are madnatory in order to clean OTEL state. # TODO use future `reset` functionality instead and save some time 💪 Given Lunar Proxy is down And Lunar Proxy is up Scenario: Request to a diagnosed endpoint is written When policies.yaml file is updated And policies.yaml includes a metrics_collector diagnosis for GET httpbinmock /anything/* requests with s3_minio as exporter And policies.yaml includes a s3_minio exporter with bucket_name: lunar-proxy-bucket and url: http://minio:9000 And policies.yaml file is saved And apply_policies command is run And A request to httpbinmock /anything/ciao is made through Lunar Proxy Then Transaction metrics are written And Traffic Metrics status should be 200 And Traffic Metrics method should be GET And Traffic Metrics normalized_url should be httpbinmock/anything/* Scenario: Request is written when global diagnosis is applied When policies.yaml file is updated And policies.yaml includes a global metrics_collector diagnosis with s3_minio as exporter And policies.yaml includes a s3_minio exporter with bucket_name: lunar-proxy-bucket and url: http://minio:9000 And policies.yaml file is saved And apply_policies command is run And A request to httpbinmock /anything/ciao is made through Lunar Proxy Then Transaction metrics are written And Traffic Metrics status should be 200 And Traffic Metrics method should be GET And Traffic Metrics normalized_url should be httpbinmock Scenario: Request is exported to Prometheus metric server When policies.yaml file is updated And policies.yaml includes a global metrics_collector diagnosis with prometheus as exporter And policies.yaml file is saved And apply_policies command is run And A request to httpbinmock /status/200 is made through Lunar Proxy And A request to httpbinmock /status/200 is made through Lunar Proxy And A request to httpbinmock /status/500 is made through Lunar Proxy Then There are 2 lunar_transaction histograms on Prometheus Metric Server And There is a histogram of status 200, normalized_url httpbinmock with 2 calls And There is a histogram of status 500, normalized_url httpbinmock with 1 calls Scenario: User-defined counters are exported to Prometheus metric server When policies.yaml file is updated And policies.yaml includes a global metrics_collector diagnosis with prometheus as exporter and custom counter for my-header response header And policies.yaml file is saved And apply_policies command is run And A request to httpbinmock /response-headers?my-header=1 is made through Lunar Proxy And A request to httpbinmock /response-headers?my-header=3 is made through Lunar Proxy And A request to httpbinmock /response-headers?my-header=non_numeric is made through Lunar Proxy And A request to httpbinmock /response-headers?my-header=2 is made through Lunar Proxy Then There are 1 lunar_transaction histograms on Prometheus Metric Server And There is a histogram of status 200, normalized_url httpbinmock with 4 calls And There is a counter named lunar_response_headers_my_header_total with the value 6