MCPX MCP Gateway

@legacy Feature: Strategy Based Throttling Remedy Scenario: Requests which exceed the limit defined by the remedy receive a rate limit error response Given API Provider is up And Lunar Proxy is up When policies.yaml file is updated And policies.yaml includes a strategy_based_throttling remedy for GET httpbinmock /headers requests with 2 requests per 1 seconds And policies.yaml file is saved And apply_policies command is run without waiting for Fluent to reload And next epoch-based 1 seconds window arrives And 3 requests are sent to httpbinmock /headers through Lunar Proxy And next epoch-based 1 seconds window arrives And 1 request is sent to httpbinmock /headers through Lunar Proxy Then Responses have 200, 200, 429, 200 status codes in order Scenario: Requests which exceed the limit with spillover Given API Provider is up And Lunar Proxy is up When policies.yaml file is updated And policies.yaml includes a strategy_based_throttling remedy for GET httpbinmock /anything requests with 2 requests per 1 seconds spillover is enabled And policies.yaml file is saved And apply_policies command is run without waiting for Fluent to reload And next epoch-based 1 seconds window arrives And 1 request is sent to httpbinmock /anything through Lunar Proxy And next epoch-based 1 seconds window arrives And 3 requests are sent to httpbinmock /anything through Lunar Proxy And 1 request is sent to httpbinmock /anything through Lunar Proxy Then Responses have 200, 200, 200, 200, 429 status codes in order @flakey Scenario: Requests which exceed the limit per endpoint defined by the remedy receive a rate limit error response Given API Provider is up And Lunar Proxy is up When policies.yaml file is updated And policies.yaml includes a strategy_based_throttling remedy for GET httpbinmock /anything/* requests with 2 requests per 1 seconds And policies.yaml includes a strategy_based_throttling remedy for GET httpbinmock /base64/{value} requests with 2 requests per 1 seconds And policies.yaml file is saved And apply_policies command is run without waiting for Fluent to reload And next epoch-based 1 seconds window arrives And 2 requests are sent to httpbinmock /anything/foo through Lunar Proxy And 1 requests are sent to httpbinmock /anything/bar through Lunar Proxy And 2 requests are sent to httpbinmock /base64/foo through Lunar Proxy And 1 requests are sent to httpbinmock /base64/bar through Lunar Proxy And next epoch-based 1 seconds window arrives And 1 request is sent to httpbinmock /anything/foo through Lunar Proxy And 1 request is sent to httpbinmock /base64/foo through Lunar Proxy Then Responses have 200, 200, 429, 200, 200, 429, 200, 200 status codes in order @flakey Scenario: Requests which exceed the quota allocation for their group defined by the remedy receive a rate limit error response Given API Provider is up And Lunar Proxy is up When policies.yaml file is updated And policies.yaml includes a strategy_based_throttling remedy for GET httpbinmock /headers requests with 4 requests per 1 seconds grouped by authorization header with quota_allocations of 25% to "123" and 75% to "456" And policies.yaml file is saved And apply_policies command is run without waiting for Fluent to reload And next epoch-based 1 seconds window arrives And 4 requests are sent to httpbinmock /headers through Lunar Proxy with authorization header set to 123 And 4 requests are sent to httpbinmock /headers through Lunar Proxy with authorization header set to 456 And next epoch-based 1 seconds window arrives And 1 request is sent to httpbinmock /headers through Lunar Proxy with authorization header set to 123 And 1 request is sent to httpbinmock /headers through Lunar Proxy with authorization header set to 456 Then Responses have 200, 429, 429, 429, 200, 200, 200, 429, 200, 200 status codes in order @flakey Scenario: Requests which exceed the quota allocation for their group defined where one group is set to 100% by the remedy receive a rate limit error response Given API Provider is up And Lunar Proxy is up When policies.yaml file is updated And policies.yaml includes a strategy_based_throttling remedy for GET httpbinmock /headers requests with 4 requests per 1 seconds grouped by authorization header with quota_allocations of 25% to "123" and 100% to "456" And policies.yaml file is saved And apply_policies command is run without waiting for Fluent to reload And next epoch-based 1 seconds window arrives And 4 requests are sent to httpbinmock /headers through Lunar Proxy with authorization header set to 123 And 5 requests are sent to httpbinmock /headers through Lunar Proxy with authorization header set to 456 And next epoch-based 1 seconds window arrives And 1 request is sent to httpbinmock /headers through Lunar Proxy with authorization header set to 123 And 2 requests are sent to httpbinmock /headers through Lunar Proxy with authorization header set to 456 Then Responses have 200, 429, 429, 429, 200, 200, 200, 200, 429, 200, 200, 200 status codes in order # The following is served as a justification for the results of the following Scenario # +----------------+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+ # | timeline (sec) | 0 | | | | | | | | | | 2 | | | # +================+=====+=====+=====+=====+=====+=====+=====+=====+=====+=====+=====+=====+=====+ # | auth value | 123 | 123 | 123 | 456 | 456 | 456 | 456 | 456 | 456 | 123 | 123 | 456 | 456 | # +----------------+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+ # | ungrouped | V | V | V | V | X | X | X | X | X | X | V | V | V | # +----------------+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+ # | grouped | V | V | X | V | V | V | V | V | V | X | V | V | V | # +----------------+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+ # | final | V | V | X | V | X | X | X | X | X | X | V | V | V | # +----------------+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+ # 200, 200, 429, 200, 429, 429, 429, 429, 429, 429, 200, 200, 200 @flakey Scenario: Requests which exceed the quota allocation with no group and then with group by the remedy receive a rate limit error response Given API Provider is up And Lunar Proxy is up When policies.yaml file is updated And policies.yaml includes a strategy_based_throttling remedy for GET httpbinmock /headers requests with 4 requests per 2 seconds And policies.yaml includes a strategy_based_throttling remedy for GET httpbinmock /headers requests with 8 requests per 4 seconds grouped by authorization header with quota_allocations of 25% to "123" and 100% to "456" And policies.yaml file is saved And apply_policies command is run without waiting for Fluent to reload And next epoch-based 4 seconds window arrives And wait 1 seconds And 3 requests are sent to httpbinmock /headers through Lunar Proxy with authorization header set to 123 And 6 requests are sent to httpbinmock /headers through Lunar Proxy with authorization header set to 456 And 1 requests are sent to httpbinmock /headers through Lunar Proxy with authorization header set to 123 And next epoch-based 4 seconds window arrives And 1 requests are sent to httpbinmock /headers through Lunar Proxy with authorization header set to 123 And 2 requests are sent to httpbinmock /headers through Lunar Proxy with authorization header set to 456 Then Responses have 200, 200, 429, 200, 429, 429, 429, 429, 429, 429, 200, 200, 200 status codes in order