Introduce grpc-downstream-authority flag

The existing `grpc-downstream-server` enabled setting a downstream server host, and via Go's gRPC default, also set the authority header. Introduce a `grpc-downstream-authority` header, allowing the user to set a string for authority, distinct from `grpc-downstream-server`. This is useful when, for example, bb needs to proxy a request through Linkerd, and still provide routing information via `authority`. Fixes #9 Signed-off-by: Andrew Seigner <[email protected]>
BuoyantIO · Aug 21, 2018 · 6b97f59 · 6b97f59
1 parent c89b095
commit 6b97f59
Show file tree

Hide file tree

Showing 5 changed files with 38 additions and 15 deletions.
diff --git a/CHANGES.md b/CHANGES.md
@@ -1,3 +1,8 @@
+## v0.0.4
+
+* Introduce `grpc-downstream-authority` flag, to enable setting authority
+  separately from `grpc-downstream-server`.
+
 ## v0.0.3
 
 * Update docs and example files to latest release.

diff --git a/README.md b/README.md
@@ -86,7 +86,6 @@ A test run using the Docker CLI should return usage information and confirm that
      $ docker run buoyantio/bb:v0.0.3
     Building Blocks or `bb` is a tool that can simulate many of the typical scenarios of a cloud-native Service-Oriented Architecture based on microservices.
 
-
     Usage:
       bb [command]
 
@@ -98,18 +97,19 @@ A test run using the Docker CLI should return usage information and confirm that
       terminus               Receives the request and returns a pre-defined response
 
     Flags:
-          --downstream-timeout duration          timeout to use when making downstream connections. (default 1m0s)
-          --fire-and-forget                      do not wait for a response when contacting downstream services.
-          --grpc-downstream-server stringSlice   list of servers (hostname:port) to send messages to using gRPC, can be repeated
-          --grpc-server-port int                 port to bind a gRPC server to (default -1)
-          --h1-downstream-server stringSlice     list of servers (protocol://hostname:port) to send messages to using HTTP 1.1, can be repeated
-          --h1-server-port int                   port to bind a HTTP 1.1 server to (default -1)
-      -h, --help                                 help for bb
-          --id string                            identifier for this container
-          --log-level string                     log level, must be one of: panic, fatal, error, warn, info, debug (default "debug")
-          --percent-failure int                  percentage of requests that this service will automatically fail
-          --sleep-in-millis int                  amount of milliseconds to wait before actually start processing a request
-          --terminate-after int                  terminate the process after this many requests
+          --downstream-timeout duration             timeout to use when making downstream connections. (default 1m0s)
+          --fire-and-forget                         do not wait for a response when contacting downstream services.
+          --grpc-downstream-authority stringSlice   list of authority headers to specify routing, if set, ordering and count should match grpc-downstream-server
+          --grpc-downstream-server stringSlice      list of servers (hostname:port) to send messages to using gRPC, can be repeated
+          --grpc-server-port int                    port to bind a gRPC server to (default -1)
+          --h1-downstream-server stringSlice        list of servers (protocol://hostname:port) to send messages to using HTTP 1.1, can be repeated
+          --h1-server-port int                      port to bind a HTTP 1.1 server to (default -1)
+      -h, --help                                    help for bb
+          --id string                               identifier for this container
+          --log-level string                        log level, must be one of: panic, fatal, error, warn, info, debug (default "debug")
+          --percent-failure int                     percentage of requests that this service will automatically fail
+          --sleep-in-millis int                     amount of milliseconds to wait before actually start processing a request
+          --terminate-after int                     terminate the process after this many requests
 
     Use "bb [command] --help" for more information about a command.
 

diff --git a/cmd/root.go b/cmd/root.go
@@ -51,6 +51,7 @@ func init() {
 	RootCmd.PersistentFlags().IntVar(&config.TerminateAfter, "terminate-after", 0, "terminate the process after this many requests")
 	RootCmd.PersistentFlags().BoolVar(&config.FireAndForget, "fire-and-forget", false, "do not wait for a response when contacting downstream services.")
 	RootCmd.PersistentFlags().StringSliceVar(&config.GRPCDownstreamServers, "grpc-downstream-server", []string{}, "list of servers (hostname:port) to send messages to using gRPC, can be repeated")
+	RootCmd.PersistentFlags().StringSliceVar(&config.GRPCDownstreamAuthorities, "grpc-downstream-authority", []string{}, "list of authority headers to specify routing, if set, ordering and count should match grpc-downstream-server")
 	RootCmd.PersistentFlags().StringSliceVar(&config.H1DownstreamServers, "h1-downstream-server", []string{}, "list of servers (protocol://hostname:port) to send messages to using HTTP 1.1, can be repeated")
 	RootCmd.PersistentFlags().DurationVar(&config.DownstreamConnectionTimeout, "downstream-timeout", time.Minute*1, "timeout to use when making downstream connections.")
 	RootCmd.PersistentFlags().StringVar(&logLevel, "log-level", log.DebugLevel.String(), "log level, must be one of: panic, fatal, error, warn, info, debug")

diff --git a/protocols/grpc.go b/protocols/grpc.go
@@ -81,8 +81,24 @@ func NewGrpcServerIfConfigured(config *service.Config, serviceHandler *service.R
 // downstream service
 func NewGrpcClientsIfConfigured(config *service.Config) ([]service.Client, error) {
 	clients := make([]service.Client, 0)
-	for _, serverURL := range config.GRPCDownstreamServers {
-		conn, err := grpc.Dial(serverURL, grpc.WithInsecure())
+
+	withAuthorities := false
+	if len(config.GRPCDownstreamAuthorities) > 0 {
+		if len(config.GRPCDownstreamAuthorities) != len(config.GRPCDownstreamServers) {
+			err := fmt.Errorf("Authorities count (%d) does not match gRPC downstream server count (%d)", len(config.GRPCDownstreamAuthorities), len(config.GRPCDownstreamServers))
+			log.Error(err)
+			return nil, err
+		}
+
+		withAuthorities = true
+	}
+
+	for i, serverURL := range config.GRPCDownstreamServers {
+		authority := ""
+		if withAuthorities {
+			authority = config.GRPCDownstreamAuthorities[i]
+		}
+		conn, err := grpc.Dial(serverURL, grpc.WithInsecure(), grpc.WithAuthority(authority))
 		if err != nil {
 			return nil, err
 		}

diff --git a/service/service.go b/service/service.go
@@ -16,6 +16,7 @@ type Config struct {
 	GRPCServerPort              int
 	H1ServerPort                int
 	GRPCDownstreamServers       []string
+	GRPCDownstreamAuthorities   []string
 	H1DownstreamServers         []string
 	PercentageFailedRequests    int
 	SleepInMillis               int