diff --git a/packages/datadog-plugin-google-cloud-pubsub/src/producer.js b/packages/datadog-plugin-google-cloud-pubsub/src/producer.js
index 94644e4e063..902b215cce1 100644
--- a/packages/datadog-plugin-google-cloud-pubsub/src/producer.js
+++ b/packages/datadog-plugin-google-cloud-pubsub/src/producer.js
@@ -9,35 +9,147 @@ class GoogleCloudPubsubProducerPlugin extends ProducerPlugin {
 
   bindStart (ctx) {
     const { request, api, projectId } = ctx
-
     if (api !== 'publish') return
 
     const messages = request.messages || []
     const topic = request.topic
-    const span = this.startSpan({ // TODO: rename
+    const messageCount = messages.length
+    const hasTraceContext = messages[0]?.attributes?.['x-datadog-trace-id']
+
+    // Collect span links from messages 2-N (skip first - it becomes parent)
+    const spanLinkData = hasTraceContext
+      ? messages.slice(1).map(msg => this._extractSpanLink(msg.attributes)).filter(Boolean)
+      : []
+
+    // Extract parent from first message
+    const firstAttrs = messages[0]?.attributes
+    const parentData = firstAttrs?.['x-datadog-trace-id'] && firstAttrs['x-datadog-parent-id']
+      ? {
+          traceId: firstAttrs['x-datadog-trace-id'],
+          spanId: firstAttrs['x-datadog-parent-id'],
+          traceIdUpper: firstAttrs['_dd.p.tid'],
+          samplingPriority: firstAttrs['x-datadog-sampling-priority']
+        }
+      : null
+
+    // Create pubsub.request span
+    const batchSpan = this.startSpan({
+      childOf: parentData ? this._extractParentContext(parentData) : undefined,
       resource: `${api} ${topic}`,
       meta: {
         'gcloud.project_id': projectId,
-        'pubsub.method': api, // TODO: remove
-        'pubsub.topic': topic
+        'pubsub.method': api,
+        'pubsub.topic': topic,
+        'span.kind': 'producer',
+        '_dd.base_service': this.tracer._service,
+        '_dd.serviceoverride.type': 'integration',
+        'pubsub.linked_message_count': spanLinkData.length || undefined,
+        operation: messageCount > 1 ? 'batched.pubsub.request' : 'pubsub.request'
+      },
+      metrics: {
+        'pubsub.batch.message_count': messageCount,
+        'pubsub.batch': messageCount > 1 ? true : undefined
       }
     }, ctx)
 
-    for (const msg of messages) {
-      if (!msg.attributes) {
-        msg.attributes = {}
+    const spanCtx = batchSpan.context()
+    const batchTraceId = spanCtx.toTraceId()
+    const batchSpanId = spanCtx.toSpanId()
+    const batchTraceIdUpper = spanCtx._trace.tags['_dd.p.tid']
+
+    // Convert to hex for storage (simpler, used directly by span links)
+    const batchTraceIdHex = BigInt(batchTraceId).toString(16).padStart(16, '0')
+    const batchSpanIdHex = BigInt(batchSpanId).toString(16).padStart(16, '0')
+
+    // Add span links as metadata
+    if (spanLinkData.length) {
+      batchSpan.setTag('_dd.span_links', JSON.stringify(
+        spanLinkData.map(link => ({
+          trace_id: link.traceId,
+          span_id: link.spanId,
+          flags: link.samplingPriority || 0
+        }))
+      ))
+    }
+
+    // Add metadata to all messages
+    messages.forEach((msg, i) => {
+      msg.attributes = msg.attributes || {}
+
+      if (!hasTraceContext) {
+        this.tracer.inject(batchSpan, 'text_map', msg.attributes)
+      }
+
+      Object.assign(msg.attributes, {
+        '_dd.pubsub_request.trace_id': batchTraceIdHex,
+        '_dd.pubsub_request.span_id': batchSpanIdHex,
+        '_dd.batch.size': String(messageCount),
+        '_dd.batch.index': String(i),
+        'gcloud.project_id': projectId,
+        'pubsub.topic': topic
+      })
+
+      if (batchTraceIdUpper) {
+        msg.attributes['_dd.pubsub_request.p.tid'] = batchTraceIdUpper
       }
-      this.tracer.inject(span, 'text_map', msg.attributes)
+
+      msg.attributes['x-dd-publish-start-time'] ??= String(Date.now())
+
       if (this.config.dsmEnabled) {
-        const payloadSize = getHeadersSize(msg)
-        const dataStreamsContext = this.tracer
-          .setCheckpoint(['direction:out', `topic:${topic}`, 'type:google-pubsub'], span, payloadSize)
+        const dataStreamsContext = this.tracer.setCheckpoint(
+          ['direction:out', `topic:${topic}`, 'type:google-pubsub'],
+          batchSpan,
+          getHeadersSize(msg)
+        )
         DsmPathwayCodec.encode(dataStreamsContext, msg.attributes)
       }
-    }
+    })
 
+    ctx.batchSpan = batchSpan
     return ctx.currentStore
   }
+
+  bindFinish (ctx) {
+    if (ctx.batchSpan && !ctx.batchSpan._duration) ctx.batchSpan.finish()
+    return super.bindFinish(ctx)
+  }
+
+  bindError (ctx) {
+    if (ctx.error && ctx.batchSpan) {
+      ctx.batchSpan.setTag('error', ctx.error)
+      ctx.batchSpan.finish()
+    }
+    return super.bindError(ctx)
+  }
+
+  _extractSpanLink (attrs) {
+    if (!attrs?.['x-datadog-trace-id'] || !attrs['x-datadog-parent-id']) return null
+
+    const lowerHex = BigInt(attrs['x-datadog-trace-id']).toString(16).padStart(16, '0')
+    const spanIdHex = BigInt(attrs['x-datadog-parent-id']).toString(16).padStart(16, '0')
+    const traceIdHex = attrs['_dd.p.tid']
+      ? attrs['_dd.p.tid'] + lowerHex
+      : lowerHex.padStart(32, '0')
+
+    return {
+      traceId: traceIdHex,
+      spanId: spanIdHex,
+      samplingPriority: attrs['x-datadog-sampling-priority']
+        ? Number.parseInt(attrs['x-datadog-sampling-priority'], 10)
+        : undefined
+    }
+  }
+
+  _extractParentContext (data) {
+    const carrier = {
+      'x-datadog-trace-id': data.traceId,
+      'x-datadog-parent-id': data.spanId
+    }
+    if (data.traceIdUpper) carrier['_dd.p.tid'] = data.traceIdUpper
+    if (data.samplingPriority) carrier['x-datadog-sampling-priority'] = String(data.samplingPriority)
+
+    return this.tracer.extract('text_map', carrier)
+  }
 }
 
 module.exports = GoogleCloudPubsubProducerPlugin