node, chain: Add extensible compression support for RPC requests

fubhy · claude · fubhy · commit a2eba9c46724 · 2025-07-22T19:57:42.000+02:00
- Replace boolean compression_enabled with Compression enum (None, Gzip) - Support per-provider compression configuration via "compression" field - Add placeholders for future compression methods (Brotli, Deflate) - Update transport layer to handle compression enum with match statement - Add comprehensive unit tests for compression configuration parsing - Update example configuration and documentation Configuration examples: compression = "gzip" # Enable gzip compression compression = "none" # Disable compression (default) Addresses issue #5671 with future-extensible design. 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>
diff --git a/Cargo.lock b/Cargo.lock
diff --git a/chain/ethereum/src/transport.rs b/chain/ethereum/src/transport.rs
@@ -1,5 +1,6 @@
 use graph::components::network_provider::ProviderName;
 use graph::endpoint::{EndpointMetrics, RequestLabels};
+use graph_node::config::Compression;
 use jsonrpc_core::types::Call;
 use jsonrpc_core::Value;
 
@@ -54,12 +55,25 @@ impl Transport {
         headers: graph::http::HeaderMap,
         metrics: Arc<EndpointMetrics>,
         provider: impl AsRef<str>,
+        compression: Compression,
     ) -> Self {
         // Unwrap: This only fails if something is wrong with the system's TLS config.
-        let client = reqwest::Client::builder()
-            .default_headers(headers)
-            .build()
-            .unwrap();
+        let mut client_builder = reqwest::Client::builder().default_headers(headers);
+
+        match compression {
+            Compression::Gzip => {
+                // Enable gzip compression/decompression for requests and responses
+                client_builder = client_builder.gzip(true);
+            }
+            Compression::None => {
+                // No compression
+            } // Future compression methods can be handled here:
+              // Compression::Brotli => {
+              //     client_builder = client_builder.brotli(true);
+              // }
+        }
+
+        let client = client_builder.build().unwrap();
 
         Transport::RPC {
             client: http::Http::with_client(client, rpc),
diff --git a/graph/Cargo.toml b/graph/Cargo.toml
@@ -26,7 +26,7 @@ chrono = "0.4.41"
 envconfig = "0.11.0"
 Inflector = "0.11.3"
 atty = "0.2"
-reqwest = { version = "0.12.15", features = ["json", "stream", "multipart"] }
+reqwest = { version = "0.12.15", features = ["json", "stream", "multipart", "gzip"] }
 ethabi = "17.2"
 hex = "0.4.3"
 http0 = { version = "0", package = "http" }
diff --git a/node/claude/plans/gzip-compression-plan.md b/node/claude/plans/gzip-compression-plan.md
@@ -0,0 +1,98 @@
+# Plan: Implement Extensible Compression for RPC Requests
+
+## Overview
+Add extensible compression support for Graph Node's outgoing RPC requests to upstream providers, configurable on a per-provider basis with future compression methods in mind.
+
+## Implementation Steps (COMPLETED)
+
+### 1. ✅ Create Compression Enum (`node/src/config.rs`)
+- Added `Compression` enum with `None` and `Gzip` variants
+- Commented placeholders for future compression methods (Brotli, Deflate)
+- Default implementation returns `Compression::None`
+
+### 2. ✅ Update Configuration Structure (`node/src/config.rs`)
+- Replaced `compression_enabled: bool` with `compression: Compression` field in `Web3Provider` struct
+- Updated all existing code to use new enum
+- Added unit tests for both "gzip" and "none" compression options
+
+### 3. ✅ Modify HTTP Transport (`chain/ethereum/src/transport.rs`)
+- Updated `Transport::new_rpc()` to accept `Compression` enum parameter
+- Implemented match statement for different compression types
+- Added comments showing where future compression methods can be added
+- Uses reqwest's `.gzip(true)` for automatic compression/decompression
+
+### 4. ✅ Update Transport Creation (`node/src/chain.rs`)
+- Pass compression enum from config to transport
+- Updated logging to show compression method using debug format
+
+### 5. ✅ Update Dependencies (`graph/Cargo.toml`)
+- Added "gzip" feature to reqwest dependency
+
+### 6. ✅ Update Test Configuration
+- Updated `full_config.toml` example to use new enum format
+- Added comprehensive unit tests for compression parsing
+
+## Configuration Examples
+
+### Gzip Compression
+```toml
+[chains.mainnet]
+provider = [
+  { 
+    label = "mainnet-rpc", 
+    details = { 
+      type = "web3", 
+      url = "http://rpc.example.com", 
+      features = ["archive"],
+      compression = "gzip"
+    }
+  }
+]
+```
+
+### No Compression (Default)
+```toml
+[chains.mainnet]
+provider = [
+  { 
+    label = "mainnet-rpc", 
+    details = { 
+      type = "web3", 
+      url = "http://rpc.example.com", 
+      features = ["archive"],
+      compression = "none"  # or omit entirely
+    }
+  }
+]
+```
+
+### Future Extension Example
+```rust
+// Future compression methods can be easily added:
+#[derive(Copy, Clone, Debug, Deserialize, Serialize, PartialEq)]
+pub enum Compression {
+    #[serde(rename = "none")]
+    None,
+    #[serde(rename = "gzip")]
+    Gzip,
+    #[serde(rename = "brotli")]
+    Brotli,
+    #[serde(rename = "deflate")]
+    Deflate,
+}
+
+// And handled in transport:
+match compression {
+    Compression::Gzip => client_builder = client_builder.gzip(true),
+    Compression::Brotli => client_builder = client_builder.brotli(true),
+    Compression::Deflate => client_builder = client_builder.deflate(true),
+    Compression::None => {} // No compression
+}
+```
+
+## Benefits of This Implementation
+- **Extensible**: Easy to add new compression methods without breaking changes
+- **Backward Compatible**: Defaults to no compression, existing configs work unchanged
+- **Type Safe**: Enum prevents invalid compression method strings
+- **Future Proof**: Clear pattern for adding Brotli, Deflate, etc.
+- **Per-Provider**: Each RPC provider can have different compression settings
diff --git a/node/resources/tests/full_config.toml b/node/resources/tests/full_config.toml
@@ -48,6 +48,7 @@ shard = "primary"
 provider = [
   { label = "mainnet-0", url = "http://rpc.mainnet.io", features = ["archive", "traces"] },
   { label = "mainnet-1", details = { type = "web3call", url = "http://rpc.mainnet.io", features = ["archive", "traces"] }},
+  { label = "mainnet-2", details = { type = "web3", url = "http://rpc.mainnet.io", features = ["archive"], compression = "gzip" }},
   { label = "firehose", details = { type = "firehose", url = "http://localhost:9000", features = [] }},
   { label = "substreams", details = { type = "substreams", url = "http://localhost:9000", features = [] }},
 ]
diff --git a/node/src/chain.rs b/node/src/chain.rs
@@ -282,7 +282,8 @@ pub async fn create_ethereum_networks_for_chain(
             logger,
             "Creating transport";
             "url" => &web3.url,
-            "capabilities" => capabilities
+            "capabilities" => capabilities,
+            "compression" => ?web3.compression
         );
 
         use crate::config::Transport::*;
@@ -293,6 +294,7 @@ pub async fn create_ethereum_networks_for_chain(
                 web3.headers.clone(),
                 endpoint_metrics.cheap_clone(),
                 &provider.label,
+                web3.compression,
             ),
             Ipc => Transport::new_ipc(&web3.url).await,
             Ws => Transport::new_ws(&web3.url).await,
diff --git a/node/src/config.rs b/node/src/config.rs
@@ -502,6 +502,7 @@ impl ChainSection {
                         features,
                         headers: Default::default(),
                         rules: vec![],
+                        compression: Compression::None,
                     }),
                 };
                 let entry = chains.entry(name.to_string()).or_insert_with(|| Chain {
@@ -705,6 +706,10 @@ pub struct Web3Provider {
 
     #[serde(default, rename = "match")]
     rules: Vec<Web3Rule>,
+
+    /// Compression method for RPC requests and responses
+    #[serde(default)]
+    pub compression: Compression,
 }
 
 impl Web3Provider {
@@ -901,6 +906,7 @@ impl<'de> Deserialize<'de> for Provider {
                             .ok_or_else(|| serde::de::Error::missing_field("features"))?,
                         headers: headers.unwrap_or_else(HeaderMap::new),
                         rules: nodes,
+                        compression: Compression::None,
                     }),
                 };
 
@@ -944,6 +950,25 @@ pub enum Transport {
     Ipc,
 }
 
+#[derive(Copy, Clone, Debug, Deserialize, Serialize, PartialEq)]
+pub enum Compression {
+    #[serde(rename = "none")]
+    None,
+    #[serde(rename = "gzip")]
+    Gzip,
+    // Future compression methods can be added here:
+    // #[serde(rename = "brotli")]
+    // Brotli,
+    // #[serde(rename = "deflate")]
+    // Deflate,
+}
+
+impl Default for Compression {
+    fn default() -> Self {
+        Compression::None
+    }
+}
+
 impl Default for Transport {
     fn default() -> Self {
         Self::Rpc
@@ -1307,6 +1332,7 @@ mod tests {
                     features: BTreeSet::new(),
                     headers: HeaderMap::new(),
                     rules: Vec::new(),
+                    compression: Compression::None,
                 }),
             },
             actual
@@ -1333,6 +1359,7 @@ mod tests {
                     features: BTreeSet::new(),
                     headers: HeaderMap::new(),
                     rules: Vec::new(),
+                    compression: Compression::None,
                 }),
             },
             actual
@@ -1440,6 +1467,7 @@ mod tests {
                     features,
                     headers,
                     rules: Vec::new(),
+                    compression: Compression::None,
                 }),
             },
             actual
@@ -1465,6 +1493,7 @@ mod tests {
                     features: BTreeSet::new(),
                     headers: HeaderMap::new(),
                     rules: Vec::new(),
+                    compression: Compression::None,
                 }),
             },
             actual
@@ -1834,6 +1863,7 @@ mod tests {
                     features: BTreeSet::new(),
                     headers: HeaderMap::new(),
                     rules: Vec::new(),
+                    compression: Compression::None,
                 }),
             },
             actual
@@ -1846,6 +1876,66 @@ mod tests {
         assert!(SubgraphLimit::Limit(10) > SubgraphLimit::Disabled);
     }
 
+    #[test]
+    fn it_parses_web3_provider_with_compression() {
+        let actual = toml::from_str(
+            r#"
+            label = "compressed"
+            details = { type = "web3", url = "http://localhost:8545", features = ["archive"], compression = "gzip" }
+        "#,
+        )
+        .unwrap();
+
+        assert_eq!(
+            Provider {
+                label: "compressed".to_owned(),
+                details: ProviderDetails::Web3(Web3Provider {
+                    transport: Transport::Rpc,
+                    url: "http://localhost:8545".to_owned(),
+                    features: {
+                        let mut features = BTreeSet::new();
+                        features.insert("archive".to_string());
+                        features
+                    },
+                    headers: HeaderMap::new(),
+                    rules: Vec::new(),
+                    compression: Compression::Gzip,
+                }),
+            },
+            actual
+        );
+    }
+
+    #[test]
+    fn it_parses_web3_provider_with_no_compression() {
+        let actual = toml::from_str(
+            r#"
+            label = "uncompressed"
+            details = { type = "web3", url = "http://localhost:8545", features = ["archive"], compression = "none" }
+        "#,
+        )
+        .unwrap();
+
+        assert_eq!(
+            Provider {
+                label: "uncompressed".to_owned(),
+                details: ProviderDetails::Web3(Web3Provider {
+                    transport: Transport::Rpc,
+                    url: "http://localhost:8545".to_owned(),
+                    features: {
+                        let mut features = BTreeSet::new();
+                        features.insert("archive".to_string());
+                        features
+                    },
+                    headers: HeaderMap::new(),
+                    rules: Vec::new(),
+                    compression: Compression::None,
+                }),
+            },
+            actual
+        );
+    }
+
     #[test]
     fn duplicated_labels_are_not_allowed_within_chain() {
         let mut actual = toml::from_str::<ChainSection>(

Original file line number	Diff line number	Diff line change
`@@ -48,6 +48,7 @@ shard = "primary"`
`48`	`48`	`provider = [`
`49`	`49`	`{ label = "mainnet-0", url = "http://rpc.mainnet.io", features = ["archive", "traces"] },`
`50`	`50`	`{ label = "mainnet-1", details = { type = "web3call", url = "http://rpc.mainnet.io", features = ["archive", "traces"] }},`
	`51`	`+ { label = "mainnet-2", details = { type = "web3", url = "http://rpc.mainnet.io", features = ["archive"], compression = "gzip" }},`
`51`	`52`	`{ label = "firehose", details = { type = "firehose", url = "http://localhost:9000", features = [] }},`
`52`	`53`	`{ label = "substreams", details = { type = "substreams", url = "http://localhost:9000", features = [] }},`
`53`	`54`	`]`