diff --git a/docs/en/changes/changes.md b/docs/en/changes/changes.md
index 8e5a10c1df13..08992c160031 100644
--- a/docs/en/changes/changes.md
+++ b/docs/en/changes/changes.md
@@ -11,6 +11,7 @@
 * Fix missing parentheses around OR conditions in `JDBCZipkinQueryDAO.getTraces()`, which caused the table filter to be bypassed for all but the first trace ID. Replaced with a proper `IN` clause.
 * Fix missing `and` keyword in `JDBCEBPFProfilingTaskDAO.getTaskRecord()` SQL query, which caused a syntax error on every invocation.
 * Fix duplicate `TABLE_COLUMN` condition in `JDBCMetadataQueryDAO.findEndpoint()`, which was binding the same parameter twice due to a copy-paste error.
+* Support MCP (Model Context Protocol) observability for Envoy AI Gateway: MCP metrics (request CPM/latency, method breakdown, backend breakdown, initialization latency, capabilities), MCP access log sampling (errors only), `ai_route_type` searchable log tag, and MCP dashboard tabs.
 
 #### UI
 
diff --git a/docs/en/setup/backend/backend-envoy-ai-gateway-monitoring.md b/docs/en/setup/backend/backend-envoy-ai-gateway-monitoring.md
index e8620edb9d8b..bd614f6f61b3 100644
--- a/docs/en/setup/backend/backend-envoy-ai-gateway-monitoring.md
+++ b/docs/en/setup/backend/backend-envoy-ai-gateway-monitoring.md
@@ -4,8 +4,9 @@
 
 [Envoy AI Gateway](https://aigateway.envoyproxy.io/) is a gateway/proxy for AI/LLM API traffic
 (OpenAI, Anthropic, AWS Bedrock, Azure OpenAI, Google Gemini, etc.) built on top of Envoy Proxy.
-It natively emits GenAI metrics and access logs via OTLP, following
-[OpenTelemetry GenAI Semantic Conventions](https://opentelemetry.io/docs/specs/semconv/gen-ai/).
+It natively emits GenAI metrics following
+[OpenTelemetry GenAI Semantic Conventions](https://opentelemetry.io/docs/specs/semconv/gen-ai/),
+and also emits MCP (Model Context Protocol) metrics and access logs via OTLP.
 
 SkyWalking receives OTLP metrics and logs directly on its gRPC port (11800) — no OpenTelemetry
 Collector is needed between the AI Gateway and SkyWalking OAP.
@@ -15,7 +16,7 @@ Collector is needed between the AI Gateway and SkyWalking OAP.
   [Envoy AI Gateway getting started](https://aigateway.envoyproxy.io/docs/getting-started/) for installation.
 
 ### Data flow
-1. Envoy AI Gateway processes LLM API requests and records GenAI metrics (token usage, latency, TTFT, TPOT).
+1. Envoy AI Gateway processes LLM API requests and MCP requests, recording GenAI metrics and MCP metrics.
 2. The AI Gateway pushes metrics and access logs via OTLP gRPC to SkyWalking OAP.
 3. SkyWalking OAP parses metrics with [MAL](../../concepts-and-designs/mal.md) rules and access logs
    with [LAL](../../concepts-and-designs/lal.md) rules.
@@ -27,14 +28,14 @@ in SkyWalking OAP. No OAP-side configuration is needed.
 
 Configure the AI Gateway to push OTLP to SkyWalking by setting these environment variables:
 
-| Env Var | Value | Purpose |
-|---------|-------|---------|
-| `OTEL_SERVICE_NAME` | Per-deployment gateway name (e.g., `my-ai-gateway`) | SkyWalking service name |
-| `OTEL_EXPORTER_OTLP_ENDPOINT` | `http://skywalking-oap:11800` | SkyWalking OAP gRPC receiver |
-| `OTEL_EXPORTER_OTLP_PROTOCOL` | `grpc` | OTLP transport |
-| `OTEL_METRICS_EXPORTER` | `otlp` | Enable OTLP metrics push |
-| `OTEL_LOGS_EXPORTER` | `otlp` | Enable OTLP access log push |
-| `OTEL_RESOURCE_ATTRIBUTES` | See below | Routing + instance + layer |
+| Env Var                       | Value                                               | Purpose                      |
+|-------------------------------|-----------------------------------------------------|------------------------------|
+| `OTEL_SERVICE_NAME`           | Per-deployment gateway name (e.g., `my-ai-gateway`) | SkyWalking service name      |
+| `OTEL_EXPORTER_OTLP_ENDPOINT` | `http://skywalking-oap:11800`                       | SkyWalking OAP gRPC receiver |
+| `OTEL_EXPORTER_OTLP_PROTOCOL` | `grpc`                                              | OTLP transport               |
+| `OTEL_METRICS_EXPORTER`       | `otlp`                                              | Enable OTLP metrics push     |
+| `OTEL_LOGS_EXPORTER`          | `otlp`                                              | Enable OTLP access log push  |
+| `OTEL_RESOURCE_ATTRIBUTES`    | See below                                           | Routing + instance + layer   |
 
 **Required resource attributes** (in `OTEL_RESOURCE_ATTRIBUTES`):
 - `job_name=envoy-ai-gateway` — Fixed routing tag for MAL/LAL rules. Same for all AI Gateway deployments.
@@ -58,47 +59,86 @@ is a service, each pod is an instance. Metrics include per-provider and per-mode
 
 #### Service Metrics
 
-| Monitoring Panel | Unit | Metric Name | Description |
-|---|---|---|---|
-| Request CPM | calls/min | meter_envoy_ai_gw_request_cpm | Requests per minute |
-| Request Latency Avg | ms | meter_envoy_ai_gw_request_latency_avg | Average request duration |
-| Request Latency Percentile | ms | meter_envoy_ai_gw_request_latency_percentile | P50/P75/P90/P95/P99 |
-| Input Token Rate | tokens/min | meter_envoy_ai_gw_input_token_rate | Input (prompt) tokens per minute |
-| Output Token Rate | tokens/min | meter_envoy_ai_gw_output_token_rate | Output (completion) tokens per minute |
-| TTFT Avg | ms | meter_envoy_ai_gw_ttft_avg | Time to First Token (streaming only) |
-| TTFT Percentile | ms | meter_envoy_ai_gw_ttft_percentile | P50/P75/P90/P95/P99 TTFT |
-| TPOT Avg | ms | meter_envoy_ai_gw_tpot_avg | Time Per Output Token (streaming only) |
-| TPOT Percentile | ms | meter_envoy_ai_gw_tpot_percentile | P50/P75/P90/P95/P99 TPOT |
+| Monitoring Panel           | Unit       | Metric Name                                  | Description                            |
+|----------------------------|------------|----------------------------------------------|----------------------------------------|
+| Request CPM                | calls/min  | meter_envoy_ai_gw_request_cpm                | Requests per minute                    |
+| Request Latency Avg        | ms         | meter_envoy_ai_gw_request_latency_avg        | Average request duration               |
+| Request Latency Percentile | ms         | meter_envoy_ai_gw_request_latency_percentile | P50/P75/P90/P95/P99                    |
+| Input Token Rate           | tokens/min | meter_envoy_ai_gw_input_token_rate           | Input (prompt) tokens per minute       |
+| Output Token Rate          | tokens/min | meter_envoy_ai_gw_output_token_rate          | Output (completion) tokens per minute  |
+| TTFT Avg                   | ms         | meter_envoy_ai_gw_ttft_avg                   | Time to First Token (streaming only)   |
+| TTFT Percentile            | ms         | meter_envoy_ai_gw_ttft_percentile            | P50/P75/P90/P95/P99 TTFT               |
+| TPOT Avg                   | ms         | meter_envoy_ai_gw_tpot_avg                   | Time Per Output Token (streaming only) |
+| TPOT Percentile            | ms         | meter_envoy_ai_gw_tpot_percentile            | P50/P75/P90/P95/P99 TPOT               |
 
 #### Provider Breakdown Metrics
 
-| Monitoring Panel | Unit | Metric Name | Description |
-|---|---|---|---|
-| Provider Request CPM | calls/min | meter_envoy_ai_gw_provider_request_cpm | Requests by provider |
-| Provider Token Rate | tokens/min | meter_envoy_ai_gw_provider_token_rate | Token rate by provider |
-| Provider Latency Avg | ms | meter_envoy_ai_gw_provider_latency_avg | Latency by provider |
+| Monitoring Panel     | Unit       | Metric Name                            | Description            |
+|----------------------|------------|----------------------------------------|------------------------|
+| Provider Request CPM | calls/min  | meter_envoy_ai_gw_provider_request_cpm | Requests by provider   |
+| Provider Token Rate  | tokens/min | meter_envoy_ai_gw_provider_token_rate  | Token rate by provider |
+| Provider Latency Avg | ms         | meter_envoy_ai_gw_provider_latency_avg | Latency by provider    |
 
 #### Model Breakdown Metrics
 
-| Monitoring Panel | Unit | Metric Name | Description |
-|---|---|---|---|
-| Model Request CPM | calls/min | meter_envoy_ai_gw_model_request_cpm | Requests by model |
-| Model Token Rate | tokens/min | meter_envoy_ai_gw_model_token_rate | Token rate by model |
-| Model Latency Avg | ms | meter_envoy_ai_gw_model_latency_avg | Latency by model |
-| Model TTFT Avg | ms | meter_envoy_ai_gw_model_ttft_avg | TTFT by model |
-| Model TPOT Avg | ms | meter_envoy_ai_gw_model_tpot_avg | TPOT by model |
+| Monitoring Panel  | Unit       | Metric Name                         | Description         |
+|-------------------|------------|-------------------------------------|---------------------|
+| Model Request CPM | calls/min  | meter_envoy_ai_gw_model_request_cpm | Requests by model   |
+| Model Token Rate  | tokens/min | meter_envoy_ai_gw_model_token_rate  | Token rate by model |
+| Model Latency Avg | ms         | meter_envoy_ai_gw_model_latency_avg | Latency by model    |
+| Model TTFT Avg    | ms         | meter_envoy_ai_gw_model_ttft_avg    | TTFT by model       |
+| Model TPOT Avg    | ms         | meter_envoy_ai_gw_model_tpot_avg    | TPOT by model       |
 
 #### Instance Metrics
 
 All service-level metrics are also available per instance (pod) with `meter_envoy_ai_gw_instance_` prefix,
 including per-provider and per-model breakdowns.
 
+### MCP Metrics
+
+When the AI Gateway is configured with MCP (Model Context Protocol) routes, SkyWalking collects
+MCP-specific metrics. These appear in the **MCP** tab on the service and instance dashboards.
+
+#### MCP Service Metrics
+
+| Monitoring Panel                      | Unit      | Metric Name                                             | Description                                                       |
+|---------------------------------------|-----------|---------------------------------------------------------|-------------------------------------------------------------------|
+| MCP Request CPM                       | calls/min | meter_envoy_ai_gw_mcp_request_cpm                       | MCP requests per minute                                           |
+| MCP Request Latency Avg               | ms        | meter_envoy_ai_gw_mcp_request_latency_avg               | Average MCP request duration                                      |
+| MCP Request Latency Percentile        | ms        | meter_envoy_ai_gw_mcp_request_latency_percentile        | P50/P75/P90/P95/P99                                               |
+| MCP Method CPM                        | calls/min | meter_envoy_ai_gw_mcp_method_cpm                        | Requests by MCP method (initialize, tools/list, tools/call, etc.) |
+| MCP Error CPM                         | calls/min | meter_envoy_ai_gw_mcp_error_cpm                         | MCP error requests per minute                                     |
+| MCP Initialization Latency Avg        | ms        | meter_envoy_ai_gw_mcp_initialization_latency_avg        | Average MCP session initialization time                           |
+| MCP Initialization Latency Percentile | ms        | meter_envoy_ai_gw_mcp_initialization_latency_percentile | P50/P75/P90/P95/P99                                               |
+| MCP Capabilities CPM                  | calls/min | meter_envoy_ai_gw_mcp_capabilities_cpm                  | Capabilities negotiated by type                                   |
+
+#### MCP Backend Breakdown Metrics
+
+| Monitoring Panel         | Unit      | Metric Name                                              | Description                    |
+|--------------------------|-----------|----------------------------------------------------------|--------------------------------|
+| Backend Request CPM      | calls/min | meter_envoy_ai_gw_mcp_backend_request_cpm                | Requests by MCP backend        |
+| Backend Latency Avg      | ms        | meter_envoy_ai_gw_mcp_backend_request_latency_avg        | Latency by MCP backend         |
+| Backend Method CPM       | calls/min | meter_envoy_ai_gw_mcp_backend_method_cpm                 | Requests by backend and method |
+| Backend Error CPM        | calls/min | meter_envoy_ai_gw_mcp_backend_error_cpm                  | Errors by MCP backend          |
+| Backend Init Latency Avg | ms        | meter_envoy_ai_gw_mcp_backend_initialization_latency_avg | Init latency by backend        |
+
+#### MCP Instance Metrics
+
+All MCP service-level metrics are also available per instance with `meter_envoy_ai_gw_mcp_instance_` prefix.
+
 ### Access Log Sampling
 
-The LAL rules apply a sampling policy to reduce storage:
+Access logs are tagged with `ai_route_type` (`llm` or `mcp`) for filtering in the log query UI.
+The `ai_route_type` tag is searchable by default.
+
+**LLM route logs:**
 - **Error responses** (HTTP status >= 400) — always persisted.
 - **Upstream failures** — always persisted.
 - **High token cost** (>= 10,000 total tokens) — persisted for cost anomaly detection.
 - Normal successful responses with low token counts are dropped.
 
-The token threshold can be adjusted in `lal/envoy-ai-gateway.yaml`.
+**MCP route logs:**
+- **Error responses** (HTTP status >= 400) — always persisted.
+- Normal MCP requests are dropped (MCP observability is covered by metrics).
+
+The sampling policy can be adjusted in `lal/envoy-ai-gateway.yaml`.
diff --git a/oap-server/server-starter/src/main/resources/application.yml b/oap-server/server-starter/src/main/resources/application.yml
index b79ff36a95c8..a0112f34d3d0 100644
--- a/oap-server/server-starter/src/main/resources/application.yml
+++ b/oap-server/server-starter/src/main/resources/application.yml
@@ -118,7 +118,7 @@ core:
     searchableTracesTags: ${SW_SEARCHABLE_TAG_KEYS:http.method,http.status_code,rpc.status_code,db.type,db.instance,mq.queue,mq.topic,mq.broker}
     # Define the set of log tag keys, which should be searchable through the GraphQL.
     # The max length of key=value should be less than 256 or will be dropped.
-    searchableLogsTags: ${SW_SEARCHABLE_LOGS_TAG_KEYS:level,http.status_code}
+    searchableLogsTags: ${SW_SEARCHABLE_LOGS_TAG_KEYS:level,http.status_code,ai_route_type}
     # Define the set of alarm tag keys, which should be searchable through the GraphQL.
     # The max length of key=value should be less than 256 or will be dropped.
     searchableAlarmTags: ${SW_SEARCHABLE_ALARM_TAG_KEYS:level}
diff --git a/oap-server/server-starter/src/main/resources/lal/envoy-ai-gateway.yaml b/oap-server/server-starter/src/main/resources/lal/envoy-ai-gateway.yaml
index 0af60377b5b1..ef6e8a518605 100644
--- a/oap-server/server-starter/src/main/resources/lal/envoy-ai-gateway.yaml
+++ b/oap-server/server-starter/src/main/resources/lal/envoy-ai-gateway.yaml
@@ -15,29 +15,52 @@
 
 # Envoy AI Gateway access log processing via OTLP.
 #
-# Sampling policy: only persist abnormal or expensive requests.
-# Normal 200 responses with low token count and no upstream failure are dropped.
+# Two rules: one for LLM route logs, one for MCP route logs.
+# LLM sampling: persist error responses (>= 400), upstream failures, or high-token requests (>= 10000).
+# MCP sampling: only persist error responses (>= 400).
+# Both tag ai_route_type for searchable filtering in the UI.
 
 rules:
-  - name: envoy-ai-gateway-access-log
+  - name: envoy-ai-gateway-llm-access-log
     layer: ENVOY_AI_GATEWAY
     dsl: |
       filter {
-        // Drop normal logs: response < 400, no upstream failure, low token count
+        // Only process LLM route logs (gen_ai.request.model is always set for LLM routes, even on errors)
+        if (tag("gen_ai.request.model") == "" || tag("gen_ai.request.model") == "-") {
+          abort {}
+        }
+
+        // Keep: error responses (>= 400), upstream failures, or high-token requests (>= 10000 total tokens)
+        // Abort logs without response_code unless there is an upstream failure
+        if (tag("response_code") == "" || tag("response_code") == "-") {
+          if (tag("upstream_transport_failure_reason") == "" || tag("upstream_transport_failure_reason") == "-") {
+            abort {}
+          }
+        }
+        // For normal responses (< 400), check upstream failure and token cost
         if (tag("response_code") != "" && tag("response_code") != "-") {
           if (tag("response_code") as Integer < 400) {
+            if (tag("upstream_transport_failure_reason") != "" && tag("upstream_transport_failure_reason") != "-") {
+              // upstream failure — keep
+            }
             if (tag("upstream_transport_failure_reason") == "" || tag("upstream_transport_failure_reason") == "-") {
+              // no upstream failure — check token cost
               if (tag("gen_ai.usage.input_tokens") != "" && tag("gen_ai.usage.input_tokens") != "-"
                   && tag("gen_ai.usage.output_tokens") != "" && tag("gen_ai.usage.output_tokens") != "-") {
                 if ((tag("gen_ai.usage.input_tokens") as Integer) + (tag("gen_ai.usage.output_tokens") as Integer) < 10000) {
                   abort {}
                 }
               }
+              if (tag("gen_ai.usage.input_tokens") == "" || tag("gen_ai.usage.input_tokens") == "-"
+                  || tag("gen_ai.usage.output_tokens") == "" || tag("gen_ai.usage.output_tokens") == "-") {
+                abort {}
+              }
             }
           }
         }
 
         extractor {
+          tag 'ai_route_type': "llm"
           tag 'gen_ai.request.model': tag("gen_ai.request.model")
           tag 'gen_ai.response.model': tag("gen_ai.response.model")
           tag 'gen_ai.provider.name': tag("gen_ai.provider.name")
@@ -50,3 +73,33 @@ rules:
         sink {
         }
       }
+
+  - name: envoy-ai-gateway-mcp-access-log
+    layer: ENVOY_AI_GATEWAY
+    dsl: |
+      filter {
+        // Only process MCP route logs
+        if (tag("mcp.method.name") == "" || tag("mcp.method.name") == "-") {
+          abort {}
+        }
+
+        // Only persist error responses (>= 400)
+        if (tag("response_code") == "" || tag("response_code") == "-") {
+          abort {}
+        }
+        if (tag("response_code") as Integer < 400) {
+          abort {}
+        }
+
+        extractor {
+          tag 'ai_route_type': "mcp"
+          tag 'mcp.method.name': tag("mcp.method.name")
+          tag 'mcp.provider.name': tag("mcp.provider.name")
+          tag 'mcp.session.id': tag("mcp.session.id")
+          tag 'response_code': tag("response_code")
+          tag 'duration': tag("duration")
+        }
+
+        sink {
+        }
+      }
diff --git a/oap-server/server-starter/src/main/resources/otel-rules/envoy-ai-gateway/gateway-mcp-instance.yaml b/oap-server/server-starter/src/main/resources/otel-rules/envoy-ai-gateway/gateway-mcp-instance.yaml
new file mode 100644
index 000000000000..8ed1d2527864
--- /dev/null
+++ b/oap-server/server-starter/src/main/resources/otel-rules/envoy-ai-gateway/gateway-mcp-instance.yaml
@@ -0,0 +1,80 @@
+# Licensed to the Apache Software Foundation (ASF) under one or more
+# contributor license agreements.  See the NOTICE file distributed with
+# this work for additional information regarding copyright ownership.
+# The ASF licenses this file to You under the Apache License, Version 2.0
+# (the "License"); you may not use this file except in compliance with
+# the License.  You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+# Envoy AI Gateway — MCP Instance-level (per-pod) metrics
+#
+# Same metrics as gateway-mcp-service.yaml but scoped to individual pods.
+# All durations are in seconds from the AI Gateway; multiply by 1000 for ms display.
+
+filter: "{ tags -> tags.job_name == 'envoy-ai-gateway' }"
+expSuffix: instance(['service_name'], ['service_instance_id'], Layer.ENVOY_AI_GATEWAY)
+metricPrefix: meter_envoy_ai_gw_mcp_instance
+
+metricsRules:
+  # ===================== Aggregate MCP metrics =====================
+
+  # MCP request CPM
+  - name: request_cpm
+    exp: mcp_request_duration_count.sum(['service_name', 'service_instance_id']).increase('PT1M')
+
+  # MCP request latency average (ms)
+  - name: request_latency_avg
+    exp: mcp_request_duration_sum.sum(['service_name', 'service_instance_id']).increase('PT1M') / mcp_request_duration_count.sum(['service_name', 'service_instance_id']).increase('PT1M') * 1000
+
+  # MCP request latency percentile (ms)
+  - name: request_latency_percentile
+    exp: mcp_request_duration.sum(['le', 'service_name', 'service_instance_id']).increase('PT1M').histogram().histogram_percentile([50,75,90,95,99]) * 1000
+
+  # MCP method invocation CPM — labeled by mcp_method_name
+  - name: method_cpm
+    exp: mcp_method_count.sum(['mcp_method_name', 'service_name', 'service_instance_id']).increase('PT1M')
+
+  # MCP error CPM
+  - name: error_cpm
+    exp: mcp_method_count.tagEqual('status', 'error').sum(['service_name', 'service_instance_id']).increase('PT1M')
+
+  # MCP initialization latency average (ms)
+  - name: initialization_latency_avg
+    exp: mcp_initialization_duration_sum.sum(['service_name', 'service_instance_id']).increase('PT1M') / mcp_initialization_duration_count.sum(['service_name', 'service_instance_id']).increase('PT1M') * 1000
+
+  # MCP initialization latency percentile (ms)
+  - name: initialization_latency_percentile
+    exp: mcp_initialization_duration.sum(['le', 'service_name', 'service_instance_id']).increase('PT1M').histogram().histogram_percentile([50,75,90,95,99]) * 1000
+
+  # MCP capabilities negotiated CPM — labeled by capability_type
+  - name: capabilities_cpm
+    exp: mcp_capabilities_negotiated.sum(['capability_type', 'service_name', 'service_instance_id']).increase('PT1M')
+
+  # ===================== Per-backend breakdown =====================
+
+  # Backend request CPM
+  - name: backend_request_cpm
+    exp: mcp_request_duration_count.sum(['mcp_backend', 'service_name', 'service_instance_id']).increase('PT1M')
+
+  # Backend request latency average (ms)
+  - name: backend_request_latency_avg
+    exp: mcp_request_duration_sum.sum(['mcp_backend', 'service_name', 'service_instance_id']).increase('PT1M') / mcp_request_duration_count.sum(['mcp_backend', 'service_name', 'service_instance_id']).increase('PT1M') * 1000
+
+  # Backend method CPM
+  - name: backend_method_cpm
+    exp: mcp_method_count.sum(['mcp_backend', 'mcp_method_name', 'service_name', 'service_instance_id']).increase('PT1M')
+
+  # Backend error CPM
+  - name: backend_error_cpm
+    exp: mcp_method_count.tagEqual('status', 'error').sum(['mcp_backend', 'service_name', 'service_instance_id']).increase('PT1M')
+
+  # Backend initialization latency average (ms)
+  - name: backend_initialization_latency_avg
+    exp: mcp_initialization_duration_sum.sum(['mcp_backend', 'service_name', 'service_instance_id']).increase('PT1M') / mcp_initialization_duration_count.sum(['mcp_backend', 'service_name', 'service_instance_id']).increase('PT1M') * 1000
diff --git a/oap-server/server-starter/src/main/resources/otel-rules/envoy-ai-gateway/gateway-mcp-service.yaml b/oap-server/server-starter/src/main/resources/otel-rules/envoy-ai-gateway/gateway-mcp-service.yaml
new file mode 100644
index 000000000000..482b8edaa305
--- /dev/null
+++ b/oap-server/server-starter/src/main/resources/otel-rules/envoy-ai-gateway/gateway-mcp-service.yaml
@@ -0,0 +1,85 @@
+# Licensed to the Apache Software Foundation (ASF) under one or more
+# contributor license agreements.  See the NOTICE file distributed with
+# this work for additional information regarding copyright ownership.
+# The ASF licenses this file to You under the Apache License, Version 2.0
+# (the "License"); you may not use this file except in compliance with
+# the License.  You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+# Envoy AI Gateway — MCP Service-level metrics
+#
+# Source OTLP metrics (dots → underscores by OTel receiver):
+#   mcp_request_duration         — Histogram (Cumulative), unit: seconds, labels: mcp_backend, error_type (on error)
+#   mcp_method_count             — Counter (Cumulative), labels: mcp_backend, mcp_method_name, status
+#   mcp_initialization_duration  — Histogram (Cumulative), unit: seconds, labels: mcp_backend
+#   mcp_capabilities_negotiated  — Counter (Cumulative), labels: mcp_backend, capability_type, capability_side
+#
+# All durations are in seconds from the AI Gateway; multiply by 1000 for ms display.
+
+filter: "{ tags -> tags.job_name == 'envoy-ai-gateway' }"
+expSuffix: service(['service_name'], Layer.ENVOY_AI_GATEWAY)
+metricPrefix: meter_envoy_ai_gw_mcp
+
+metricsRules:
+  # ===================== Aggregate MCP metrics =====================
+
+  # MCP request CPM — count of all MCP requests per minute
+  - name: request_cpm
+    exp: mcp_request_duration_count.sum(['service_name']).increase('PT1M')
+
+  # MCP request latency average (ms)
+  - name: request_latency_avg
+    exp: mcp_request_duration_sum.sum(['service_name']).increase('PT1M') / mcp_request_duration_count.sum(['service_name']).increase('PT1M') * 1000
+
+  # MCP request latency percentile (ms)
+  - name: request_latency_percentile
+    exp: mcp_request_duration.sum(['le', 'service_name']).increase('PT1M').histogram().histogram_percentile([50,75,90,95,99]) * 1000
+
+  # MCP method invocation CPM — labeled by mcp_method_name
+  - name: method_cpm
+    exp: mcp_method_count.sum(['mcp_method_name', 'service_name']).increase('PT1M')
+
+  # MCP error CPM — only error status
+  - name: error_cpm
+    exp: mcp_method_count.tagEqual('status', 'error').sum(['service_name']).increase('PT1M')
+
+  # MCP initialization latency average (ms)
+  - name: initialization_latency_avg
+    exp: mcp_initialization_duration_sum.sum(['service_name']).increase('PT1M') / mcp_initialization_duration_count.sum(['service_name']).increase('PT1M') * 1000
+
+  # MCP initialization latency percentile (ms)
+  - name: initialization_latency_percentile
+    exp: mcp_initialization_duration.sum(['le', 'service_name']).increase('PT1M').histogram().histogram_percentile([50,75,90,95,99]) * 1000
+
+  # MCP capabilities negotiated CPM — labeled by capability_type
+  - name: capabilities_cpm
+    exp: mcp_capabilities_negotiated.sum(['capability_type', 'service_name']).increase('PT1M')
+
+  # ===================== Per-backend breakdown =====================
+
+  # Backend request CPM — labeled by mcp_backend
+  - name: backend_request_cpm
+    exp: mcp_request_duration_count.sum(['mcp_backend', 'service_name']).increase('PT1M')
+
+  # Backend request latency average (ms) — labeled by mcp_backend
+  - name: backend_request_latency_avg
+    exp: mcp_request_duration_sum.sum(['mcp_backend', 'service_name']).increase('PT1M') / mcp_request_duration_count.sum(['mcp_backend', 'service_name']).increase('PT1M') * 1000
+
+  # Backend method CPM — labeled by mcp_backend and mcp_method_name
+  - name: backend_method_cpm
+    exp: mcp_method_count.sum(['mcp_backend', 'mcp_method_name', 'service_name']).increase('PT1M')
+
+  # Backend error CPM — labeled by mcp_backend
+  - name: backend_error_cpm
+    exp: mcp_method_count.tagEqual('status', 'error').sum(['mcp_backend', 'service_name']).increase('PT1M')
+
+  # Backend initialization latency average (ms) — labeled by mcp_backend
+  - name: backend_initialization_latency_avg
+    exp: mcp_initialization_duration_sum.sum(['mcp_backend', 'service_name']).increase('PT1M') / mcp_initialization_duration_count.sum(['mcp_backend', 'service_name']).increase('PT1M') * 1000
diff --git a/oap-server/server-starter/src/main/resources/ui-initialized-templates/envoy_ai_gateway/envoy-ai-gateway-instance.json b/oap-server/server-starter/src/main/resources/ui-initialized-templates/envoy_ai_gateway/envoy-ai-gateway-instance.json
index fe314a11c000..a70b4a2b9de3 100644
--- a/oap-server/server-starter/src/main/resources/ui-initialized-templates/envoy_ai_gateway/envoy-ai-gateway-instance.json
+++ b/oap-server/server-starter/src/main/resources/ui-initialized-templates/envoy_ai_gateway/envoy-ai-gateway-instance.json
@@ -36,8 +36,8 @@
                     }
                   ],
                   "widget": {
-                    "title": "Request CPM",
-                    "tips": "Calls Per Minute — total requests through this pod"
+                    "title": "Request CPM (calls / min)",
+                    "tips": "Total requests through this pod"
                   }
                 },
                 {
@@ -62,7 +62,7 @@
                     }
                   ],
                   "widget": {
-                    "title": "Request Latency Avg"
+                    "title": "Request Latency Avg (ms)"
                   }
                 },
                 {
@@ -87,7 +87,7 @@
                     }
                   ],
                   "widget": {
-                    "title": "Request Latency Percentile",
+                    "title": "Request Latency Percentile (ms)",
                     "tips": "P50 / P75 / P90 / P95 / P99"
                   }
                 },
@@ -113,7 +113,7 @@
                     }
                   ],
                   "widget": {
-                    "title": "Input Token Rate",
+                    "title": "Input Token Rate (tokens / min)",
                     "tips": "Input (prompt) tokens per minute sent to LLM providers"
                   }
                 },
@@ -139,7 +139,7 @@
                     }
                   ],
                   "widget": {
-                    "title": "Output Token Rate",
+                    "title": "Output Token Rate (tokens / min)",
                     "tips": "Output (completion) tokens per minute generated by LLM providers"
                   }
                 },
@@ -165,7 +165,7 @@
                     }
                   ],
                   "widget": {
-                    "title": "Time to First Token Avg (TTFT)",
+                    "title": "Time to First Token Avg - TTFT (ms)",
                     "tips": "Average time to first token for streaming requests"
                   }
                 },
@@ -191,7 +191,7 @@
                     }
                   ],
                   "widget": {
-                    "title": "Time to First Token Percentile (TTFT)",
+                    "title": "Time to First Token Percentile - TTFT (ms)",
                     "tips": "P50 / P75 / P90 / P95 / P99"
                   }
                 },
@@ -217,7 +217,7 @@
                     }
                   ],
                   "widget": {
-                    "title": "Time Per Output Token Avg (TPOT)",
+                    "title": "Time Per Output Token Avg - TPOT (ms)",
                     "tips": "Average inter-token latency for streaming requests"
                   }
                 },
@@ -243,7 +243,7 @@
                     }
                   ],
                   "widget": {
-                    "title": "Time Per Output Token Percentile (TPOT)",
+                    "title": "Time Per Output Token Percentile - TPOT (ms)",
                     "tips": "P50 / P75 / P90 / P95 / P99"
                   }
                 }
@@ -274,7 +274,7 @@
                     }
                   ],
                   "widget": {
-                    "title": "Request CPM by Provider"
+                    "title": "Request CPM by Provider (calls / min)"
                   }
                 },
                 {
@@ -299,7 +299,7 @@
                     }
                   ],
                   "widget": {
-                    "title": "Token Rate by Provider"
+                    "title": "Token Rate by Provider (tokens / min)"
                   }
                 },
                 {
@@ -324,7 +324,7 @@
                     }
                   ],
                   "widget": {
-                    "title": "Latency Avg by Provider"
+                    "title": "Latency Avg by Provider (ms)"
                   }
                 }
               ]
@@ -354,7 +354,7 @@
                     }
                   ],
                   "widget": {
-                    "title": "Request CPM by Model"
+                    "title": "Request CPM by Model (calls / min)"
                   }
                 },
                 {
@@ -379,7 +379,7 @@
                     }
                   ],
                   "widget": {
-                    "title": "Token Rate by Model"
+                    "title": "Token Rate by Model (tokens / min)"
                   }
                 },
                 {
@@ -404,7 +404,7 @@
                     }
                   ],
                   "widget": {
-                    "title": "Latency Avg by Model"
+                    "title": "Latency Avg by Model (ms)"
                   }
                 },
                 {
@@ -429,7 +429,7 @@
                     }
                   ],
                   "widget": {
-                    "title": "Time to First Token Avg by Model (TTFT)"
+                    "title": "Time to First Token Avg by Model - TTFT (ms)"
                   }
                 },
                 {
@@ -454,7 +454,339 @@
                     }
                   ],
                   "widget": {
-                    "title": "Time Per Output Token Avg by Model (TPOT)"
+                    "title": "Time Per Output Token Avg by Model - TPOT (ms)"
+                  }
+                }
+              ]
+            },
+            {
+              "name": "MCP",
+              "children": [
+                {
+                  "x": 0,
+                  "y": 0,
+                  "w": 8,
+                  "h": 13,
+                  "i": "0",
+                  "type": "Widget",
+                  "expressions": [
+                    "meter_envoy_ai_gw_mcp_instance_request_cpm"
+                  ],
+                  "graph": {
+                    "type": "Line",
+                    "showXAxis": true,
+                    "showYAxis": true
+                  },
+                  "metricConfig": [
+                    {
+                      "label": "MCP Request CPM",
+                      "unit": "calls/min"
+                    }
+                  ],
+                  "widget": {
+                    "title": "MCP Request CPM (calls / min)"
+                  }
+                },
+                {
+                  "x": 8,
+                  "y": 0,
+                  "w": 8,
+                  "h": 13,
+                  "i": "1",
+                  "type": "Widget",
+                  "expressions": [
+                    "meter_envoy_ai_gw_mcp_instance_request_latency_avg"
+                  ],
+                  "graph": {
+                    "type": "Line",
+                    "showXAxis": true,
+                    "showYAxis": true
+                  },
+                  "metricConfig": [
+                    {
+                      "label": "Avg Latency",
+                      "unit": "ms"
+                    }
+                  ],
+                  "widget": {
+                    "title": "MCP Request Latency Avg (ms)"
+                  }
+                },
+                {
+                  "x": 16,
+                  "y": 0,
+                  "w": 8,
+                  "h": 13,
+                  "i": "2",
+                  "type": "Widget",
+                  "expressions": [
+                    "meter_envoy_ai_gw_mcp_instance_request_latency_percentile"
+                  ],
+                  "graph": {
+                    "type": "Line",
+                    "showXAxis": true,
+                    "showYAxis": true
+                  },
+                  "metricConfig": [
+                    {
+                      "label": "Latency Percentile",
+                      "unit": "ms"
+                    }
+                  ],
+                  "widget": {
+                    "title": "MCP Request Latency Percentile (ms)",
+                    "tips": "P50 / P75 / P90 / P95 / P99"
+                  }
+                },
+                {
+                  "x": 0,
+                  "y": 13,
+                  "w": 8,
+                  "h": 13,
+                  "i": "3",
+                  "type": "Widget",
+                  "expressions": [
+                    "meter_envoy_ai_gw_mcp_instance_error_cpm"
+                  ],
+                  "graph": {
+                    "type": "Line",
+                    "showXAxis": true,
+                    "showYAxis": true
+                  },
+                  "metricConfig": [
+                    {
+                      "label": "Error CPM",
+                      "unit": "calls/min"
+                    }
+                  ],
+                  "widget": {
+                    "title": "MCP Error CPM (calls / min)"
+                  }
+                },
+                {
+                  "x": 8,
+                  "y": 13,
+                  "w": 8,
+                  "h": 13,
+                  "i": "4",
+                  "type": "Widget",
+                  "expressions": [
+                    "aggregate_labels(meter_envoy_ai_gw_mcp_instance_method_cpm,sum(mcp_method_name))"
+                  ],
+                  "graph": {
+                    "type": "Line",
+                    "showXAxis": true,
+                    "showYAxis": true
+                  },
+                  "metricConfig": [
+                    {
+                      "label": "Method CPM",
+                      "unit": "calls/min"
+                    }
+                  ],
+                  "widget": {
+                    "title": "MCP Method CPM (calls / min)"
+                  }
+                },
+                {
+                  "x": 16,
+                  "y": 13,
+                  "w": 8,
+                  "h": 13,
+                  "i": "5",
+                  "type": "Widget",
+                  "expressions": [
+                    "meter_envoy_ai_gw_mcp_instance_initialization_latency_avg"
+                  ],
+                  "graph": {
+                    "type": "Line",
+                    "showXAxis": true,
+                    "showYAxis": true
+                  },
+                  "metricConfig": [
+                    {
+                      "label": "Init Latency",
+                      "unit": "ms"
+                    }
+                  ],
+                  "widget": {
+                    "title": "MCP Initialization Latency Avg (ms)"
+                  }
+                },
+                {
+                  "x": 0,
+                  "y": 26,
+                  "w": 8,
+                  "h": 13,
+                  "i": "6",
+                  "type": "Widget",
+                  "expressions": [
+                    "aggregate_labels(meter_envoy_ai_gw_mcp_instance_backend_request_cpm,sum(mcp_backend))"
+                  ],
+                  "graph": {
+                    "type": "Line",
+                    "showXAxis": true,
+                    "showYAxis": true
+                  },
+                  "metricConfig": [
+                    {
+                      "label": "Backend CPM",
+                      "unit": "calls/min"
+                    }
+                  ],
+                  "widget": {
+                    "title": "MCP Request CPM by Backend (calls / min)"
+                  }
+                },
+                {
+                  "x": 8,
+                  "y": 26,
+                  "w": 8,
+                  "h": 13,
+                  "i": "7",
+                  "type": "Widget",
+                  "expressions": [
+                    "aggregate_labels(meter_envoy_ai_gw_mcp_instance_backend_request_latency_avg,avg(mcp_backend))"
+                  ],
+                  "graph": {
+                    "type": "Line",
+                    "showXAxis": true,
+                    "showYAxis": true
+                  },
+                  "metricConfig": [
+                    {
+                      "label": "Backend Latency",
+                      "unit": "ms"
+                    }
+                  ],
+                  "widget": {
+                    "title": "MCP Latency Avg by Backend (ms)"
+                  }
+                },
+                {
+                  "x": 16,
+                  "y": 26,
+                  "w": 8,
+                  "h": 13,
+                  "i": "8",
+                  "type": "Widget",
+                  "expressions": [
+                    "aggregate_labels(meter_envoy_ai_gw_mcp_instance_backend_error_cpm,sum(mcp_backend))"
+                  ],
+                  "graph": {
+                    "type": "Line",
+                    "showXAxis": true,
+                    "showYAxis": true
+                  },
+                  "metricConfig": [
+                    {
+                      "label": "Backend Error CPM",
+                      "unit": "calls/min"
+                    }
+                  ],
+                  "widget": {
+                    "title": "MCP Error CPM by Backend (calls / min)"
+                  }
+                },
+                {
+                  "x": 0,
+                  "y": 39,
+                  "w": 8,
+                  "h": 13,
+                  "i": "9",
+                  "type": "Widget",
+                  "expressions": [
+                    "meter_envoy_ai_gw_mcp_instance_initialization_latency_percentile"
+                  ],
+                  "graph": {
+                    "type": "Line",
+                    "showXAxis": true,
+                    "showYAxis": true
+                  },
+                  "metricConfig": [
+                    {
+                      "label": "Init Latency Percentile",
+                      "unit": "ms"
+                    }
+                  ],
+                  "widget": {
+                    "title": "MCP Initialization Latency Percentile (ms)",
+                    "tips": "P50 / P75 / P90 / P95 / P99"
+                  }
+                },
+                {
+                  "x": 8,
+                  "y": 39,
+                  "w": 8,
+                  "h": 13,
+                  "i": "10",
+                  "type": "Widget",
+                  "expressions": [
+                    "aggregate_labels(meter_envoy_ai_gw_mcp_instance_capabilities_cpm,sum(capability_type))"
+                  ],
+                  "graph": {
+                    "type": "Line",
+                    "showXAxis": true,
+                    "showYAxis": true
+                  },
+                  "metricConfig": [
+                    {
+                      "label": "Capabilities CPM",
+                      "unit": "calls/min"
+                    }
+                  ],
+                  "widget": {
+                    "title": "MCP Capabilities Negotiated (calls / min)"
+                  }
+                },
+                {
+                  "x": 16,
+                  "y": 39,
+                  "w": 8,
+                  "h": 13,
+                  "i": "11",
+                  "type": "Widget",
+                  "expressions": [
+                    "aggregate_labels(meter_envoy_ai_gw_mcp_instance_backend_method_cpm,sum(mcp_backend,mcp_method_name))"
+                  ],
+                  "graph": {
+                    "type": "Line",
+                    "showXAxis": true,
+                    "showYAxis": true
+                  },
+                  "metricConfig": [
+                    {
+                      "label": "Backend Method CPM",
+                      "unit": "calls/min"
+                    }
+                  ],
+                  "widget": {
+                    "title": "MCP Method CPM by Backend (calls / min)"
+                  }
+                },
+                {
+                  "x": 0,
+                  "y": 52,
+                  "w": 8,
+                  "h": 13,
+                  "i": "12",
+                  "type": "Widget",
+                  "expressions": [
+                    "aggregate_labels(meter_envoy_ai_gw_mcp_instance_backend_initialization_latency_avg,avg(mcp_backend))"
+                  ],
+                  "graph": {
+                    "type": "Line",
+                    "showXAxis": true,
+                    "showYAxis": true
+                  },
+                  "metricConfig": [
+                    {
+                      "label": "Backend Init Latency",
+                      "unit": "ms"
+                    }
+                  ],
+                  "widget": {
+                    "title": "MCP Init Latency Avg by Backend (ms)"
                   }
                 }
               ]
diff --git a/oap-server/server-starter/src/main/resources/ui-initialized-templates/envoy_ai_gateway/envoy-ai-gateway-service.json b/oap-server/server-starter/src/main/resources/ui-initialized-templates/envoy_ai_gateway/envoy-ai-gateway-service.json
index e2599eee1b0a..4ac89f5f1110 100644
--- a/oap-server/server-starter/src/main/resources/ui-initialized-templates/envoy_ai_gateway/envoy-ai-gateway-service.json
+++ b/oap-server/server-starter/src/main/resources/ui-initialized-templates/envoy_ai_gateway/envoy-ai-gateway-service.json
@@ -36,8 +36,8 @@
                     }
                   ],
                   "widget": {
-                    "title": "Request CPM",
-                    "tips": "Calls Per Minute — total requests through the AI Gateway"
+                    "title": "Request CPM (calls / min)",
+                    "tips": "Total requests through the AI Gateway"
                   }
                 },
                 {
@@ -62,7 +62,7 @@
                     }
                   ],
                   "widget": {
-                    "title": "Request Latency Avg"
+                    "title": "Request Latency Avg (ms)"
                   }
                 },
                 {
@@ -87,7 +87,7 @@
                     }
                   ],
                   "widget": {
-                    "title": "Request Latency Percentile",
+                    "title": "Request Latency Percentile (ms)",
                     "tips": "P50 / P75 / P90 / P95 / P99"
                   }
                 },
@@ -113,7 +113,7 @@
                     }
                   ],
                   "widget": {
-                    "title": "Input Token Rate",
+                    "title": "Input Token Rate (tokens / min)",
                     "tips": "Input (prompt) tokens per minute sent to LLM providers"
                   }
                 },
@@ -139,7 +139,7 @@
                     }
                   ],
                   "widget": {
-                    "title": "Output Token Rate",
+                    "title": "Output Token Rate (tokens / min)",
                     "tips": "Output (completion) tokens per minute generated by LLM providers"
                   }
                 },
@@ -165,8 +165,8 @@
                     }
                   ],
                   "widget": {
-                    "title": "Time to First Token Avg (TTFT)",
-                    "tips": "Average time to first token for streaming requests"
+                    "title": "Time to First Token Avg - TTFT (ms)",
+                    "tips": "Average TTFT for streaming requests"
                   }
                 },
                 {
@@ -191,8 +191,8 @@
                     }
                   ],
                   "widget": {
-                    "title": "Time to First Token Percentile (TTFT)",
-                    "tips": "P50 / P75 / P90 / P95 / P99"
+                    "title": "Time to First Token Percentile - TTFT (ms)",
+                    "tips": "TTFT P50 / P75 / P90 / P95 / P99"
                   }
                 },
                 {
@@ -217,8 +217,8 @@
                     }
                   ],
                   "widget": {
-                    "title": "Time Per Output Token Avg (TPOT)",
-                    "tips": "Average inter-token latency for streaming requests"
+                    "title": "Time Per Output Token Avg - TPOT (ms)",
+                    "tips": "Average TPOT inter-token latency for streaming requests"
                   }
                 },
                 {
@@ -243,8 +243,8 @@
                     }
                   ],
                   "widget": {
-                    "title": "Time Per Output Token Percentile (TPOT)",
-                    "tips": "P50 / P75 / P90 / P95 / P99"
+                    "title": "Time Per Output Token Percentile - TPOT (ms)",
+                    "tips": "TPOT P50 / P75 / P90 / P95 / P99"
                   }
                 }
               ]
@@ -274,7 +274,7 @@
                     }
                   ],
                   "widget": {
-                    "title": "Request CPM by Provider"
+                    "title": "Request CPM by Provider (calls / min)"
                   }
                 },
                 {
@@ -299,7 +299,7 @@
                     }
                   ],
                   "widget": {
-                    "title": "Token Rate by Provider"
+                    "title": "Token Rate by Provider (tokens / min)"
                   }
                 },
                 {
@@ -324,7 +324,7 @@
                     }
                   ],
                   "widget": {
-                    "title": "Latency Avg by Provider"
+                    "title": "Latency Avg by Provider (ms)"
                   }
                 }
               ]
@@ -354,7 +354,7 @@
                     }
                   ],
                   "widget": {
-                    "title": "Request CPM by Model"
+                    "title": "Request CPM by Model (calls / min)"
                   }
                 },
                 {
@@ -379,7 +379,7 @@
                     }
                   ],
                   "widget": {
-                    "title": "Token Rate by Model"
+                    "title": "Token Rate by Model (tokens / min)"
                   }
                 },
                 {
@@ -404,7 +404,7 @@
                     }
                   ],
                   "widget": {
-                    "title": "Latency Avg by Model"
+                    "title": "Latency Avg by Model (ms)"
                   }
                 },
                 {
@@ -429,7 +429,7 @@
                     }
                   ],
                   "widget": {
-                    "title": "Time to First Token Avg by Model (TTFT)"
+                    "title": "Time to First Token Avg by Model - TTFT (ms)"
                   }
                 },
                 {
@@ -454,7 +454,342 @@
                     }
                   ],
                   "widget": {
-                    "title": "Time Per Output Token Avg by Model (TPOT)"
+                    "title": "Time Per Output Token Avg by Model - TPOT (ms)"
+                  }
+                }
+              ]
+            },
+            {
+              "name": "MCP",
+              "children": [
+                {
+                  "x": 0,
+                  "y": 0,
+                  "w": 8,
+                  "h": 13,
+                  "i": "0",
+                  "type": "Widget",
+                  "expressions": [
+                    "meter_envoy_ai_gw_mcp_request_cpm"
+                  ],
+                  "graph": {
+                    "type": "Line",
+                    "showXAxis": true,
+                    "showYAxis": true
+                  },
+                  "metricConfig": [
+                    {
+                      "label": "MCP Request CPM",
+                      "unit": "calls/min"
+                    }
+                  ],
+                  "widget": {
+                    "title": "MCP Request CPM (calls / min)",
+                    "tips": "Total MCP requests per minute"
+                  }
+                },
+                {
+                  "x": 8,
+                  "y": 0,
+                  "w": 8,
+                  "h": 13,
+                  "i": "1",
+                  "type": "Widget",
+                  "expressions": [
+                    "meter_envoy_ai_gw_mcp_request_latency_avg"
+                  ],
+                  "graph": {
+                    "type": "Line",
+                    "showXAxis": true,
+                    "showYAxis": true
+                  },
+                  "metricConfig": [
+                    {
+                      "label": "Avg Latency",
+                      "unit": "ms"
+                    }
+                  ],
+                  "widget": {
+                    "title": "MCP Request Latency Avg (ms)"
+                  }
+                },
+                {
+                  "x": 16,
+                  "y": 0,
+                  "w": 8,
+                  "h": 13,
+                  "i": "2",
+                  "type": "Widget",
+                  "expressions": [
+                    "meter_envoy_ai_gw_mcp_request_latency_percentile"
+                  ],
+                  "graph": {
+                    "type": "Line",
+                    "showXAxis": true,
+                    "showYAxis": true
+                  },
+                  "metricConfig": [
+                    {
+                      "label": "Latency Percentile",
+                      "unit": "ms"
+                    }
+                  ],
+                  "widget": {
+                    "title": "MCP Request Latency Percentile (ms)",
+                    "tips": "P50 / P75 / P90 / P95 / P99"
+                  }
+                },
+                {
+                  "x": 0,
+                  "y": 13,
+                  "w": 8,
+                  "h": 13,
+                  "i": "3",
+                  "type": "Widget",
+                  "expressions": [
+                    "meter_envoy_ai_gw_mcp_error_cpm"
+                  ],
+                  "graph": {
+                    "type": "Line",
+                    "showXAxis": true,
+                    "showYAxis": true
+                  },
+                  "metricConfig": [
+                    {
+                      "label": "Error CPM",
+                      "unit": "calls/min"
+                    }
+                  ],
+                  "widget": {
+                    "title": "MCP Error CPM (calls / min)"
+                  }
+                },
+                {
+                  "x": 8,
+                  "y": 13,
+                  "w": 8,
+                  "h": 13,
+                  "i": "4",
+                  "type": "Widget",
+                  "expressions": [
+                    "aggregate_labels(meter_envoy_ai_gw_mcp_method_cpm,sum(mcp_method_name))"
+                  ],
+                  "graph": {
+                    "type": "Line",
+                    "showXAxis": true,
+                    "showYAxis": true
+                  },
+                  "metricConfig": [
+                    {
+                      "label": "Method CPM",
+                      "unit": "calls/min"
+                    }
+                  ],
+                  "widget": {
+                    "title": "MCP Method CPM (calls / min)",
+                    "tips": "Requests per minute by MCP method (initialize, tools/list, tools/call, etc.)"
+                  }
+                },
+                {
+                  "x": 16,
+                  "y": 13,
+                  "w": 8,
+                  "h": 13,
+                  "i": "5",
+                  "type": "Widget",
+                  "expressions": [
+                    "meter_envoy_ai_gw_mcp_initialization_latency_avg"
+                  ],
+                  "graph": {
+                    "type": "Line",
+                    "showXAxis": true,
+                    "showYAxis": true
+                  },
+                  "metricConfig": [
+                    {
+                      "label": "Init Latency",
+                      "unit": "ms"
+                    }
+                  ],
+                  "widget": {
+                    "title": "MCP Initialization Latency Avg (ms)",
+                    "tips": "Average time to initialize MCP session with backend"
+                  }
+                },
+                {
+                  "x": 0,
+                  "y": 26,
+                  "w": 8,
+                  "h": 13,
+                  "i": "6",
+                  "type": "Widget",
+                  "expressions": [
+                    "aggregate_labels(meter_envoy_ai_gw_mcp_backend_request_cpm,sum(mcp_backend))"
+                  ],
+                  "graph": {
+                    "type": "Line",
+                    "showXAxis": true,
+                    "showYAxis": true
+                  },
+                  "metricConfig": [
+                    {
+                      "label": "Backend CPM",
+                      "unit": "calls/min"
+                    }
+                  ],
+                  "widget": {
+                    "title": "MCP Request CPM by Backend (calls / min)"
+                  }
+                },
+                {
+                  "x": 8,
+                  "y": 26,
+                  "w": 8,
+                  "h": 13,
+                  "i": "7",
+                  "type": "Widget",
+                  "expressions": [
+                    "aggregate_labels(meter_envoy_ai_gw_mcp_backend_request_latency_avg,avg(mcp_backend))"
+                  ],
+                  "graph": {
+                    "type": "Line",
+                    "showXAxis": true,
+                    "showYAxis": true
+                  },
+                  "metricConfig": [
+                    {
+                      "label": "Backend Latency",
+                      "unit": "ms"
+                    }
+                  ],
+                  "widget": {
+                    "title": "MCP Latency Avg by Backend (ms)"
+                  }
+                },
+                {
+                  "x": 16,
+                  "y": 26,
+                  "w": 8,
+                  "h": 13,
+                  "i": "8",
+                  "type": "Widget",
+                  "expressions": [
+                    "aggregate_labels(meter_envoy_ai_gw_mcp_backend_error_cpm,sum(mcp_backend))"
+                  ],
+                  "graph": {
+                    "type": "Line",
+                    "showXAxis": true,
+                    "showYAxis": true
+                  },
+                  "metricConfig": [
+                    {
+                      "label": "Backend Error CPM",
+                      "unit": "calls/min"
+                    }
+                  ],
+                  "widget": {
+                    "title": "MCP Error CPM by Backend (calls / min)"
+                  }
+                },
+                {
+                  "x": 0,
+                  "y": 39,
+                  "w": 8,
+                  "h": 13,
+                  "i": "9",
+                  "type": "Widget",
+                  "expressions": [
+                    "meter_envoy_ai_gw_mcp_initialization_latency_percentile"
+                  ],
+                  "graph": {
+                    "type": "Line",
+                    "showXAxis": true,
+                    "showYAxis": true
+                  },
+                  "metricConfig": [
+                    {
+                      "label": "Init Latency Percentile",
+                      "unit": "ms"
+                    }
+                  ],
+                  "widget": {
+                    "title": "MCP Initialization Latency Percentile (ms)",
+                    "tips": "P50 / P75 / P90 / P95 / P99"
+                  }
+                },
+                {
+                  "x": 8,
+                  "y": 39,
+                  "w": 8,
+                  "h": 13,
+                  "i": "10",
+                  "type": "Widget",
+                  "expressions": [
+                    "aggregate_labels(meter_envoy_ai_gw_mcp_capabilities_cpm,sum(capability_type))"
+                  ],
+                  "graph": {
+                    "type": "Line",
+                    "showXAxis": true,
+                    "showYAxis": true
+                  },
+                  "metricConfig": [
+                    {
+                      "label": "Capabilities CPM",
+                      "unit": "calls/min"
+                    }
+                  ],
+                  "widget": {
+                    "title": "MCP Capabilities Negotiated (calls / min)"
+                  }
+                },
+                {
+                  "x": 16,
+                  "y": 39,
+                  "w": 8,
+                  "h": 13,
+                  "i": "11",
+                  "type": "Widget",
+                  "expressions": [
+                    "aggregate_labels(meter_envoy_ai_gw_mcp_backend_method_cpm,sum(mcp_backend,mcp_method_name))"
+                  ],
+                  "graph": {
+                    "type": "Line",
+                    "showXAxis": true,
+                    "showYAxis": true
+                  },
+                  "metricConfig": [
+                    {
+                      "label": "Backend Method CPM",
+                      "unit": "calls/min"
+                    }
+                  ],
+                  "widget": {
+                    "title": "MCP Method CPM by Backend (calls / min)"
+                  }
+                },
+                {
+                  "x": 0,
+                  "y": 52,
+                  "w": 8,
+                  "h": 13,
+                  "i": "12",
+                  "type": "Widget",
+                  "expressions": [
+                    "aggregate_labels(meter_envoy_ai_gw_mcp_backend_initialization_latency_avg,avg(mcp_backend))"
+                  ],
+                  "graph": {
+                    "type": "Line",
+                    "showXAxis": true,
+                    "showYAxis": true
+                  },
+                  "metricConfig": [
+                    {
+                      "label": "Backend Init Latency",
+                      "unit": "ms"
+                    }
+                  ],
+                  "widget": {
+                    "title": "MCP Init Latency Avg by Backend (ms)"
                   }
                 }
               ]
diff --git a/oap-server/server-starter/src/main/resources/ui-initialized-templates/virtual_genai/virtual-genai-model.json b/oap-server/server-starter/src/main/resources/ui-initialized-templates/virtual_genai/virtual-genai-model.json
index a5672eb2d64e..79ffba319a8d 100644
--- a/oap-server/server-starter/src/main/resources/ui-initialized-templates/virtual_genai/virtual-genai-model.json
+++ b/oap-server/server-starter/src/main/resources/ui-initialized-templates/virtual_genai/virtual-genai-model.json
@@ -202,8 +202,9 @@
                     "gen_ai_model_latency_percentile"
                   ],
                   "graph": {
-                    "type": "Bar",
-                    "showBackground": true
+                    "type": "Line",
+                    "showXAxis": true,
+                    "showYAxis": true
                   },
                   "widget": {
                     "name": "LatencyPercentile",
@@ -267,8 +268,9 @@
                     }
                   ],
                   "graph": {
-                    "type": "Bar",
-                    "showBackground": true
+                    "type": "Line",
+                    "showXAxis": true,
+                    "showYAxis": true
                   },
                   "id": "0-0-21",
                   "moved": false,
diff --git a/oap-server/server-starter/src/main/resources/ui-initialized-templates/virtual_genai/virtual-genai-provider.json b/oap-server/server-starter/src/main/resources/ui-initialized-templates/virtual_genai/virtual-genai-provider.json
index 0a8be506d44f..1d25ad3317d2 100644
--- a/oap-server/server-starter/src/main/resources/ui-initialized-templates/virtual_genai/virtual-genai-provider.json
+++ b/oap-server/server-starter/src/main/resources/ui-initialized-templates/virtual_genai/virtual-genai-provider.json
@@ -228,8 +228,9 @@
                     "gen_ai_provider_latency_percentile"
                   ],
                   "graph": {
-                    "type": "Bar",
-                    "showBackground": true
+                    "type": "Line",
+                    "showXAxis": true,
+                    "showYAxis": true
                   },
                   "widget": {
                     "name": "LatencyPercentile",
@@ -333,5 +334,3 @@
     }
   }
 ]
-
-
diff --git a/test/e2e-v2/cases/envoy-ai-gateway/docker-compose.yml b/test/e2e-v2/cases/envoy-ai-gateway/docker-compose.yml
index 89669cf6fd31..c2a64dc2dd6e 100644
--- a/test/e2e-v2/cases/envoy-ai-gateway/docker-compose.yml
+++ b/test/e2e-v2/cases/envoy-ai-gateway/docker-compose.yml
@@ -13,12 +13,13 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-# Envoy AI Gateway e2e — ai-gateway-cli + Ollama + SkyWalking OAP
+# Envoy AI Gateway e2e — ai-gateway-cli + Ollama + MCP Everything Server + SkyWalking OAP
 #
 # Architecture:
-#   trigger → ai-gateway-cli (port 1975) → ollama (port 11434)
-#                  ↓ OTLP gRPC
-#             oap (port 11800) → banyandb
+#   LLM trigger → ai-gateway-cli (port 1975) → ollama (port 11434)
+#   MCP trigger → ai-gateway-cli (port 1975) → mcp-server (port 3001)
+#                       ↓ OTLP gRPC
+#                  oap (port 11800) → banyandb
 
 services:
   banyandb:
@@ -52,10 +53,27 @@ services:
       timeout: 60s
       retries: 120
 
+  mcp-server:
+    image: tzolov/mcp-everything-server:v3
+    command: ["node", "dist/index.js", "streamableHttp"]
+    networks:
+      - e2e
+    expose:
+      - 3001
+    healthcheck:
+      test: ["CMD-SHELL", "wget -qO- --header='Content-Type: application/json' --header='Accept: application/json, text/event-stream' --post-data='{\"jsonrpc\":\"2.0\",\"id\":1,\"method\":\"initialize\",\"params\":{\"protocolVersion\":\"2025-03-26\",\"capabilities\":{},\"clientInfo\":{\"name\":\"healthcheck\",\"version\":\"1.0\"}}}' http://localhost:3001/mcp || exit 1"]
+      interval: 5s
+      timeout: 10s
+      retries: 30
+
   aigw:
     # TODO: pin to a release version once ai-gateway-cli HTTP listener is available in a release
     image: envoyproxy/ai-gateway-cli:latest
-    command: run --run-id=0
+    command:
+      - "run"
+      - "--run-id=0"
+      - "--mcp-json"
+      - '{"mcpServers":{"everything":{"type":"http","url":"http://mcp-server:3001/mcp"}}}'
     environment:
       OPENAI_API_KEY: "dummy-key-not-used"
       OPENAI_BASE_URL: "http://ollama:11434/v1"
@@ -71,7 +89,7 @@ services:
     networks:
       - e2e
     healthcheck:
-      test: ["CMD", "aigw", "healthcheck"]
+      test: ["CMD", "/app", "healthcheck"]
       interval: 5s
       timeout: 60s
       retries: 120
@@ -80,6 +98,8 @@ services:
         condition: service_healthy
       ollama:
         condition: service_healthy
+      mcp-server:
+        condition: service_healthy
 
 networks:
   e2e:
diff --git a/test/e2e-v2/cases/envoy-ai-gateway/e2e.yaml b/test/e2e-v2/cases/envoy-ai-gateway/e2e.yaml
index 18aaa417b839..1969c9e911b8 100644
--- a/test/e2e-v2/cases/envoy-ai-gateway/e2e.yaml
+++ b/test/e2e-v2/cases/envoy-ai-gateway/e2e.yaml
@@ -15,14 +15,15 @@
 
 # Envoy AI Gateway e2e test (docker-compose)
 #
-# Validates ENVOY_AI_GATEWAY layer metrics and logs via OTLP from ai-gateway-cli.
+# Validates ENVOY_AI_GATEWAY layer LLM + MCP metrics and logs via OTLP from ai-gateway-cli.
 #
 # Architecture:
-#   trigger (curl) → ai-gateway-cli (port 1975) → Ollama (port 11434)
-#                          ↓ OTLP gRPC
-#                     SkyWalking OAP (port 11800)
-#                          ↓
-#                     BanyanDB
+#   LLM trigger (curl) → ai-gateway-cli (port 1975) → Ollama (port 11434)
+#   MCP trigger (curl) → ai-gateway-cli (port 1975) → MCP Everything Server (port 3001)
+#                              ↓ OTLP gRPC
+#                         SkyWalking OAP (port 11800)
+#                              ↓
+#                         BanyanDB
 
 setup:
   env: compose
@@ -47,6 +48,32 @@ setup:
             -d '{"model":"nonexistent-model","messages":[{"role":"user","content":"Hi"}]}'
           sleep 1
         done
+    - name: Send MCP requests for metric verification
+      command: |
+        # Initialize MCP session and capture session ID
+        SESSION_ID=$(curl -sS -f -D- --max-time 15 \
+          http://${aigw_host}:${aigw_1975}/mcp \
+          -H 'Content-Type: application/json' \
+          -d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2025-03-26","capabilities":{},"clientInfo":{"name":"e2e-test","version":"1.0"}}}' \
+          2>&1 | grep -i '^Mcp-Session-Id:' | tr -d '\r' | sed 's/^[Mm]cp-[Ss]ession-[Ii]d: *//')
+        if [ -z "$SESSION_ID" ]; then
+          echo "Failed to capture MCP session ID" >&2
+          exit 1
+        fi
+        sleep 2
+        # tools/list
+        curl -sS -f --max-time 15 -o /dev/null \
+          http://${aigw_host}:${aigw_1975}/mcp \
+          -H 'Content-Type: application/json' \
+          -H "Mcp-Session-Id: $SESSION_ID" \
+          -d '{"jsonrpc":"2.0","id":2,"method":"tools/list","params":{}}'
+        sleep 1
+        # tools/call — call a nonexistent tool to generate an error (no -f, expects 4xx)
+        curl -sS --max-time 15 -o /dev/null \
+          http://${aigw_host}:${aigw_1975}/mcp \
+          -H 'Content-Type: application/json' \
+          -H "Mcp-Session-Id: $SESSION_ID" \
+          -d '{"jsonrpc":"2.0","id":3,"method":"tools/call","params":{"name":"nonexistent_tool","arguments":{}}}'
 
 trigger:
   action: http
diff --git a/test/e2e-v2/cases/envoy-ai-gateway/envoy-ai-gateway-cases.yaml b/test/e2e-v2/cases/envoy-ai-gateway/envoy-ai-gateway-cases.yaml
index 19f11d7dc70b..c9fd498d43b9 100644
--- a/test/e2e-v2/cases/envoy-ai-gateway/envoy-ai-gateway-cases.yaml
+++ b/test/e2e-v2/cases/envoy-ai-gateway/envoy-ai-gateway-cases.yaml
@@ -45,6 +45,26 @@ cases:
   - query: swctl --display yaml --base-url=http://${oap_host}:${oap_12800}/graphql metrics exec --expression=meter_envoy_ai_gw_model_latency_avg --service-name=e2e-ai-gateway
     expected: expected/metrics-has-value.yml
 
-  # Access logs — error requests (404) should be persisted by LAL sampling
-  - query: swctl --display yaml --base-url=http://${oap_host}:${oap_12800}/graphql logs ls --service-name=e2e-ai-gateway
+  # ===================== MCP metrics =====================
+
+  # MCP aggregate metrics
+  - query: swctl --display yaml --base-url=http://${oap_host}:${oap_12800}/graphql metrics exec --expression=meter_envoy_ai_gw_mcp_request_cpm --service-name=e2e-ai-gateway
+    expected: expected/metrics-has-value.yml
+  - query: swctl --display yaml --base-url=http://${oap_host}:${oap_12800}/graphql metrics exec --expression=meter_envoy_ai_gw_mcp_request_latency_avg --service-name=e2e-ai-gateway
+    expected: expected/metrics-has-value.yml
+  - query: swctl --display yaml --base-url=http://${oap_host}:${oap_12800}/graphql metrics exec --expression=meter_envoy_ai_gw_mcp_request_latency_percentile --service-name=e2e-ai-gateway
+    expected: expected/metrics-has-value-label.yml
+
+  # MCP method breakdown
+  - query: swctl --display yaml --base-url=http://${oap_host}:${oap_12800}/graphql metrics exec --expression=meter_envoy_ai_gw_mcp_method_cpm --service-name=e2e-ai-gateway
+    expected: expected/metrics-has-value.yml
+
+  # MCP backend breakdown
+  - query: swctl --display yaml --base-url=http://${oap_host}:${oap_12800}/graphql metrics exec --expression=meter_envoy_ai_gw_mcp_backend_request_cpm --service-name=e2e-ai-gateway
+    expected: expected/metrics-has-value.yml
+
+  # ===================== Access logs =====================
+
+  # LLM error requests (404) should be persisted with ai_route_type=llm
+  - query: swctl --display yaml --base-url=http://${oap_host}:${oap_12800}/graphql logs ls --service-name=e2e-ai-gateway --tags=ai_route_type=llm
     expected: expected/logs.yml
diff --git a/test/e2e-v2/cases/envoy-ai-gateway/expected/logs.yml b/test/e2e-v2/cases/envoy-ai-gateway/expected/logs.yml
index 37a7f56d4184..6efc226799ea 100644
--- a/test/e2e-v2/cases/envoy-ai-gateway/expected/logs.yml
+++ b/test/e2e-v2/cases/envoy-ai-gateway/expected/logs.yml
@@ -28,6 +28,8 @@ logs:
     content: {{ notEmpty .content }}
     tags:
       {{- contains .tags }}
+      - key: ai_route_type
+        value: llm
       - key: response_code
         value: "404"
       - key: gen_ai.request.model
diff --git a/test/e2e-v2/cases/storage/expected/config-dump.yml b/test/e2e-v2/cases/storage/expected/config-dump.yml
index 5dded79d67c9..59c96cb3bcd6 100644
--- a/test/e2e-v2/cases/storage/expected/config-dump.yml
+++ b/test/e2e-v2/cases/storage/expected/config-dump.yml
@@ -161,7 +161,7 @@ receiver-sharing-server.default.gRPCSslTrustedCAsPath=
 aws-firehose.default.firehoseAccessKey=******
 core.default.autocompleteTagValuesQueryMaxSize=100
 agent-analyzer.default.noUpstreamRealAddressAgents=6000,9000
-core.default.searchableLogsTags=level,http.status_code
+core.default.searchableLogsTags=level,http.status_code,ai_route_type
 core.default.role=Mixed
 receiver-sharing-server.default.gRPCSslKeyPath=
 receiver-pprof.provider=default