prototype(protobuf): extend toBinaryFast with map + oneof (+ real OTel fixture)

benchmarks/proto/nested.proto

@@ -17,18 +17,34 @@ package bench.v1;
 
 // nested.proto intentionally mirrors the shape of the OTLP trace export
 // request (opentelemetry.proto.collector.trace.v1.ExportTraceServiceRequest),
-// simplified to the fields most traffic goes through. This is a realistic
-// hotpath workload — see open-telemetry/opentelemetry-js#6221 for the
-// production incident that motivates measuring this shape.
+// now extended to exercise the full AnyValue oneof and a map<string,string>
+// so that the fast-path encoder benchmark matches a realistic OTel payload.
+// See open-telemetry/opentelemetry-js#6221 for the production incident that
+// motivates measuring this shape.
+
+// AnyValue is the oneof-typed value carried by every KeyValue attribute in
+// OTLP. We include the common leaf types; array/kvlist are intentionally
+// omitted — they nest through KeyValue again and would dominate the
+// benchmark on a distribution we don't observe in practice.
+message AnyValue {
+  oneof value {
+    string string_value = 1;
+    bool bool_value = 2;
+    int64 int_value = 3;
+    double double_value = 4;
+    bytes bytes_value = 7;
+  }
+}
 
 message KeyValue {
   string key = 1;
-  string string_value = 2;
+  AnyValue value = 2;
 }
 
 message InstrumentationScope {
   string name = 1;
   string version = 2;
+  repeated KeyValue attributes = 3;
 }
 
 message Span {
@@ -47,6 +63,11 @@ message ScopeSpans {
 
 message Resource {
   repeated KeyValue attributes = 1;
+  // `labels` exercises map<string, string> encoding on the fast path.
+  // Real OTLP payloads don't carry a labels map, but maps show up in
+  // many other ConnectRPC schemas and we want the benchmark to reflect
+  // that fast-path cost too.
+  map<string, string> labels = 2;
 }
 
 message ResourceSpans {

benchmarks/src/bench-comparison-protobufjs.ts

@@ -62,7 +62,20 @@ function buildOtelLikePayload(): Record<string, unknown> {
   for (let i = 0; i < SPAN_COUNT; i++) {
     const attributes: unknown[] = [];
     for (let j = 0; j < 10; j++) {
-      attributes.push({ key: `k${j}`, stringValue: `v${i}-${j}` });
+      // AnyValue oneof: mostly string, some int, some bool — matches the
+      // distribution the fast-path benchmark feeds into the reflective
+      // encoder via the same fixture.
+      let anyValue: Record<string, unknown>;
+      if (j === 2 || j === 5) {
+        anyValue = { value: { case: "intValue", value: BigInt(200 + j) } };
+      } else if (j === 8) {
+        anyValue = { value: { case: "boolValue", value: (i + j) % 7 === 0 } };
+      } else {
+        anyValue = {
+          value: { case: "stringValue", value: `v${i}-${j}` },
+        };
+      }
+      attributes.push({ key: `k${j}`, value: anyValue });
     }
     spans.push({
       traceId: new Uint8Array(16),
@@ -76,7 +89,14 @@ function buildOtelLikePayload(): Record<string, unknown> {
   return {
     resourceSpans: [
       {
-        resource: { attributes: [] },
+        resource: {
+          attributes: [],
+          labels: {
+            env: "production",
+            region: "us-east-1",
+            cluster: "bench-cluster",
+          },
+        },
         scopeSpans: [
           {
             scope: { name: "@example/tracer", version: "1.0.0" },
@@ -91,15 +111,39 @@ function buildOtelLikePayload(): Record<string, unknown> {
 // protobufjs uses Long for 64-bit fields by default. We generate with
 // --force-long to get consistent typing; here we still pass BigInt-like
 // plain numbers via strings so both paths encode the same timestamp.
-// Prepare a payload variant with string timestamps for pbjs, so the
-// write paths are comparable (no JSON conversion cost inside the hot loop).
+// Prepare a payload variant for pbjs with:
+//   - string timestamps (no JSON conversion cost inside the hot loop)
+//   - AnyValue oneof represented as a plain field rather than the
+//     `{ case, value }` ADT protobuf-es uses, because protobufjs stores
+//     oneof members directly on the parent message.
 function buildOtelLikePayloadForPbjs(): Record<string, unknown> {
   const base = buildOtelLikePayload();
   // biome-ignore lint/suspicious/noExplicitAny: in-place shape munging
   const resourceSpans = (base.resourceSpans as any[])[0];
   for (const span of resourceSpans.scopeSpans[0].spans) {
     span.startTimeUnixNano = "1700000000000000000";
     span.endTimeUnixNano = "1700000000000001000";
+    // biome-ignore lint/suspicious/noExplicitAny: in-place shape munging
+    for (const attr of span.attributes as any[]) {
+      const adt = attr.value?.value as
+        | { case: string; value: unknown }
+        | undefined;
+      if (adt) {
+        const pbjsKey =
+          adt.case === "intValue"
+            ? "intValue"
+            : adt.case === "boolValue"
+              ? "boolValue"
+              : "stringValue";
+        // protobufjs also accepts int64 as string without Long.
+        attr.value = {
+          [pbjsKey]:
+            adt.case === "intValue"
+              ? (adt.value as bigint).toString()
+              : adt.value,
+        };
+      }
+    }
   }
   return base;
 }

benchmarks/src/bench-create-toBinary.ts

@@ -21,6 +21,7 @@ import { Bench } from "tinybench";
 import { create, toBinary, toBinaryFast } from "@bufbuild/protobuf";
 import { SimpleMessageSchema } from "./gen/small_pb.js";
 import {
+  AnyValueSchema,
   ExportTraceRequestSchema,
   ResourceSpansSchema,
   ScopeSpansSchema,
@@ -62,7 +63,9 @@ export async function runCreateToBinaryBench() {
           attrs.push(
             create(KeyValueSchema, {
               key: `k${j}`,
-              stringValue: `v${i}-${j}`,
+              value: create(AnyValueSchema, {
+                value: { case: "stringValue", value: `v${i}-${j}` },
+              }),
             }),
           );
         }
@@ -104,7 +107,9 @@ export async function runCreateToBinaryBench() {
           attrs.push(
             create(KeyValueSchema, {
               key: `k${j}`,
-              stringValue: `v${i}-${j}`,
+              value: create(AnyValueSchema, {
+                value: { case: "stringValue", value: `v${i}-${j}` },
+              }),
             }),
           );
         }

benchmarks/src/bench-create.ts

@@ -20,6 +20,7 @@ import { Bench } from "tinybench";
 import { create } from "@bufbuild/protobuf";
 import { SimpleMessageSchema } from "./gen/small_pb.js";
 import {
+  AnyValueSchema,
   ExportTraceRequestSchema,
   ResourceSpansSchema,
   ScopeSpansSchema,
@@ -56,7 +57,9 @@ export async function runCreateBench() {
           attrs.push(
             create(KeyValueSchema, {
               key: `k${j}`,
-              stringValue: `v${i}-${j}`,
+              value: create(AnyValueSchema, {
+                value: { case: "stringValue", value: `v${i}-${j}` },
+              }),
             }),
           );
         }

benchmarks/src/bench-memory.ts

@@ -47,7 +47,19 @@ function buildInit(): Record<string, unknown> {
   for (let i = 0; i < SPAN_COUNT; i++) {
     const attributes: unknown[] = [];
     for (let j = 0; j < 10; j++) {
-      attributes.push({ key: `k${j}`, stringValue: `v${i}-${j}` });
+      // Mirror the distribution used in bench-comparison-protobufjs so the
+      // memory numbers reflect the same workload as the throughput runs.
+      let anyValue: Record<string, unknown>;
+      if (j === 2 || j === 5) {
+        anyValue = { value: { case: "intValue", value: BigInt(200 + j) } };
+      } else if (j === 8) {
+        anyValue = { value: { case: "boolValue", value: (i + j) % 7 === 0 } };
+      } else {
+        anyValue = {
+          value: { case: "stringValue", value: `v${i}-${j}` },
+        };
+      }
+      attributes.push({ key: `k${j}`, value: anyValue });
     }
     spans.push({
       traceId: new Uint8Array(16),
@@ -61,7 +73,14 @@ function buildInit(): Record<string, unknown> {
   return {
     resourceSpans: [
       {
-        resource: { attributes: [] },
+        resource: {
+          attributes: [],
+          labels: {
+            env: "production",
+            region: "us-east-1",
+            cluster: "bench-cluster",
+          },
+        },
         scopeSpans: [
           {
             scope: { name: "@example/tracer", version: "1.0.0" },
@@ -80,6 +99,26 @@ function buildInitForPbjs(): Record<string, unknown> {
   for (const span of resourceSpans.scopeSpans[0].spans) {
     span.startTimeUnixNano = "1700000000000000000";
     span.endTimeUnixNano = "1700000000000001000";
+    // biome-ignore lint/suspicious/noExplicitAny: in-place shape munging
+    for (const attr of span.attributes as any[]) {
+      const adt = attr.value?.value as
+        | { case: string; value: unknown }
+        | undefined;
+      if (adt) {
+        const pbjsKey =
+          adt.case === "intValue"
+            ? "intValue"
+            : adt.case === "boolValue"
+              ? "boolValue"
+              : "stringValue";
+        attr.value = {
+          [pbjsKey]:
+            adt.case === "intValue"
+              ? (adt.value as bigint).toString()
+              : adt.value,
+        };
+      }
+    }
   }
   return base;
 }