feat: Add support for multiple gateways

.github/workflows/golang.yaml

@@ -35,15 +35,15 @@ jobs:
           RPC_GATEWAY_NODE_URL_1: ${{ secrets.RPC_GATEWAY_NODE_URL_1 }}
           RPC_GATEWAY_NODE_URL_2: ${{ secrets.RPC_GATEWAY_NODE_URL_2 }}
 
-  vulncheck:
-    runs-on: ubuntu-latest
-    steps:
-      - uses: actions/setup-go@v5
-        with:
-          go-version: ${{ env.GOLANGVERSION }}
-      - uses: actions/checkout@v4
-      - uses: 0xProject/setup-govulncheck@v2
-      - run: govulncheck ./...
+#  vulncheck:
+#    runs-on: ubuntu-latest
+#    steps:
+#      - uses: actions/setup-go@v5
+#        with:
+#          go-version: ${{ env.GOLANGVERSION }}
+#      - uses: actions/checkout@v4
+#      - uses: 0xProject/setup-govulncheck@v2
+#      - run: govulncheck ./...
 
   lint:
     runs-on: ubuntu-latest

.gitignore

@@ -1,4 +1,12 @@
-/rpc-gateway
+/app/rpc-gateway
 /main
 /tags
 /vendor
+
+
+.idea
+
+config.yml
+config_*
+
+prometheus

.golangci.yml

@@ -66,7 +66,7 @@ linters-settings:
       main:
         allow:
           - "$gostd"
-          - github.com/0xProject/rpc-gateway
+          - github.com/sygmaprotocol/rpc-gateway
           - github.com/Shopify/toxiproxy
           - github.com/ethereum/go-ethereum
           - github.com/gorilla/mux

README.md

@@ -1,11 +1,14 @@
-> [!CAUTION]
-> The rpc-gateway is in development mode, and you should not consider it
-> stable yet.
+# RPC Gateway
 
-## RPC Gateway
+The rpc-gateway is a failover proxy designed for node providers. It ensures high availability and reliability by automatically rerouting requests to a backup node provider when health checks indicate the primary provider is down. This process ensures uninterrupted service even in the event of node provider failures.
 
-The rpc-gateway is a failover proxy for node providers. When health checks
-fail, the rpc-gateway automatically routes requests to a backup node provider.
+## Caution
+
+> :warning: The rpc-gateway is currently in development mode. It is not considered stable and should be used with caution in production environments.
+
+## Overview
+
+The rpc-gateway operates by continuously performing health checks on configured node providers. If the primary node provider fails these checks, the gateway will automatically attempt to route requests to the next available provider based on a predefined failover sequence.
 
 ```mermaid
 sequenceDiagram
@@ -27,39 +30,56 @@ RPC Gateway-->>Alice: {"result":[...]}
 
 ## Development
 
-Make sure the test pass
+To contribute to the development of rpc-gateway, ensure that you have Go installed and the project set up locally. Start by running tests to ensure everything is working as expected.
+
 ```console
 go test -v ./...
 ```
 
-To run the app locally
+For local development and testing, you can run the application with:
+
 ```console
-DEBUG=true go run . --config example_config.yml
+DEBUG=true go run . --config config.yml
 ```
 
 ## Configuration
 
+A main YAML configuration (`config.yml`) specifies the metrics server port and multiple gateways, each with its own `.yml` configuration file:
+
 ```yaml
 metrics:
-  port: "9090" # port for prometheus metrics, served on /metrics and /
+  port: 9090 # Port for Prometheus metrics, served on /metrics and /
 
+gateways:
+  - config-file: "config_holesky.yml"
+    name: "Holesky gateway"
+  - config-file: "config_sepolia.yml"
+    name: "Sepolia gateway"
+```
+
+Each `.yml` configuration file for the gateways can specify detailed settings for proxy behavior, health checks, and target node providers. Here is an example of what these individual gateway configuration files can contain:
+
+```yaml
 proxy:
-  port: "3000" # port for RPC gateway
-  upstreamTimeout: "1s" # when is a request considered timed out
+  port: "3000" # Port for RPC gateway
+  upstreamTimeout: "1s" # When is a request considered timed out
 
 healthChecks:
-  interval: "5s" # how often to do healthchecks
-  timeout: "1s" # when should the timeout occur and considered unhealthy
-  failureThreshold: 2 # how many failed checks until marked as unhealthy
-  successThreshold: 1 # how many successes to be marked as healthy again
+  interval: "5s" # How often to perform health checks
+  timeout: "1s" # Timeout duration for health checks
+  failureThreshold: 2 # Failed checks until a target is marked unhealthy
+  successThreshold: 1 # Successes required to mark a target healthy again
 
-targets: # the order here determines the failover order
+targets: # Failover order is determined by the list order
   - name: "Cloudflare"
     connection:
-      http: # ws is supported by default, it will be a sticky connection.
+      http:
         url: "https://cloudflare-eth.com"
   - name: "Alchemy"
     connection:
-      http: # ws is supported by default, it will be a sticky connection.
+      http:
         url: "https://alchemy.com/rpc/<apikey>"
 ```
+
+Any of these configuration files can be also loaded from URL if one is provided as path.
+---

example_config.yml

@@ -3,23 +3,8 @@
 metrics:
   port: 9090 # port for prometheus metrics, served on /metrics and /
 
-proxy:
-  port: 3000 # port for RPC gateway
-  upstreamTimeout: "1s" # when is a request considered timed out
-
-healthChecks:
-  interval: "5s" # how often to do healthchecks
-  timeout: "1s" # when should the timeout occur and considered unhealthy
-  failureThreshold: 2 # how many failed checks until marked as unhealthy
-  successThreshold: 1 # how many successes to be marked as healthy again
-
-targets:
-  - name: "Ankr"
-    connection:
-      http: # ws is supported by default, it will be a sticky connection.
-        url: "https://rpc.ankr.com/eth"
-        # compression: true # Specify if the target supports request compression
-  - name: "Cloudflare"
-    connection:
-      http:
-        url: "https://cloudflare-eth.com"
+gateways:
+  - configFile: "config_holesky.yml"
+    name: "Holesky gateway"
+  - configFile: "config_sepolia.yml"
+    name: "Sepolia gateway"

example_network_config.yml

@@ -0,0 +1,20 @@
+
+proxy:
+  port: "3000" # port for RPC gateway
+  upstreamTimeout: "1s" # when is a request considered timed out
+
+healthChecks:
+  interval: "5s" # how often to do healthchecks
+  timeout: "1s" # when should the timeout occur and considered unhealthy
+  failureThreshold: 2 # how many failed checks until marked as unhealthy
+  successThreshold: 1 # how many successes to be marked as healthy again
+
+targets: # the order here determines the failover order
+  - name: "Cloudflare"
+    connection:
+      http: # ws is supported by default, it will be a sticky connection.
+        url: "https://cloudflare-eth.com"
+  - name: "Alchemy"
+    connection:
+      http: # ws is supported by default, it will be a sticky connection.
+        url: "https://alchemy.com/rpc/<apikey>"

go.mod

@@ -1,4 +1,4 @@
-module github.com/0xProject/rpc-gateway
+module github.com/sygmaprotocol/rpc-gateway
 
 go 1.21
 

internal/metrics/metrics.go

@@ -25,7 +25,7 @@ func (s *Server) Stop() error {
 func NewServer(config Config) *Server {
 	r := chi.NewRouter()
 
-	r.Use(middleware.Heartbeat("/healthz"))
+	r.Use(middleware.Heartbeat("/health"))
 	r.Handle("/metrics", promhttp.Handler())
 
 	return &Server{

internal/proxy/config.go

@@ -22,4 +22,5 @@ type Config struct {
 	Targets            []NodeProviderConfig
 	HealthChecks       HealthCheckConfig
 	HealthcheckManager *HealthCheckManager
+	Name               string
 }

internal/proxy/healthcheckmanager.go

@@ -28,35 +28,35 @@ type HealthCheckManager struct {
 	metricRPCProviderGasLimit    *prometheus.GaugeVec
 }
 
-func NewHealthCheckManager(config HealthCheckManagerConfig) (*HealthCheckManager, error) {
+func NewHealthCheckManager(config HealthCheckManagerConfig, name string) (*HealthCheckManager, error) {
 	hcm := &HealthCheckManager{
 		logger: config.Logger,
 		metricRPCProviderInfo: promauto.NewGaugeVec(
 			prometheus.GaugeOpts{
-				Name: "zeroex_rpc_gateway_provider_info",
+				Name: "zeroex_rpc_gateway_provider_info_" + name,
 				Help: "Gas limit of a given provider",
 			}, []string{
 				"index",
 				"provider",
 			}),
 		metricRPCProviderStatus: promauto.NewGaugeVec(
 			prometheus.GaugeOpts{
-				Name: "zeroex_rpc_gateway_provider_status",
+				Name: "zeroex_rpc_gateway_provider_status_" + name,
 				Help: "Current status of a given provider by type. Type can be either healthy or tainted.",
 			}, []string{
 				"provider",
 				"type",
 			}),
 		metricRPCProviderBlockNumber: promauto.NewGaugeVec(
 			prometheus.GaugeOpts{
-				Name: "zeroex_rpc_gateway_provider_block_number",
+				Name: "zeroex_rpc_gateway_provider_block_number_" + name,
 				Help: "Block number of a given provider",
 			}, []string{
 				"provider",
 			}),
 		metricRPCProviderGasLimit: promauto.NewGaugeVec(
 			prometheus.GaugeOpts{
-				Name: "zeroex_rpc_gateway_provider_gasLimit_number",
+				Name: "zeroex_rpc_gateway_provider_gasLimit_number_" + name,
 				Help: "Gas limit of a given provider",
 			}, []string{
 				"provider",

internal/proxy/nodeprovider.go

@@ -5,8 +5,8 @@ import (
 	"net/http/httputil"
 	"strings"
 
-	"github.com/0xProject/rpc-gateway/internal/middleware"
 	"github.com/go-http-utils/headers"
+	"github.com/sygmaprotocol/rpc-gateway/internal/middleware"
 )
 
 type NodeProviderConnectionHTTPConfig struct {