Feature/rag metering (#75)

* Add metering-rag for RAG token metrics

* Make metering executable

* Add metering-rag to Prom metrics

* Metricate the right queue

* Update templates to include metering-rag

prometheus/prometheus.yml

@@ -30,6 +30,7 @@ scrape_configs:
         - 'kg-extract-topics:8000'
         - 'kg-extract-relationships:8000'
         - 'metering:8000'
+        - 'metering-rag:8000'
         - 'store-graph-embeddings:8000'
         - 'store-triples:8000'
         - 'text-completion:8000'

templates/components/trustgraph.jsonnet

@@ -132,5 +132,37 @@ local prompt = import "prompt-template.jsonnet";
 
     },
 
+    "metering-rag" +: {
+    
+        create:: function(engine)
+
+            local container =
+                engine.container("metering-rag")
+                    .with_image(images.trustgraph)
+                    .with_command([
+                        "metering",
+                        "-p",
+                        url.pulsar,
+                        "-i",
+                        "non-persistent://tg/response/text-completion-rag-response",
+                    ])
+                    .with_limits("0.5", "128M")
+                    .with_reservations("0.1", "128M");
+
+            local containerSet = engine.containers(
+                "metering-rag", [ container ]
+            );
+
+            local service =
+                engine.internalService(containerSet)
+                .with_port(8000, 8000, "metrics");
+
+            engine.resources([
+                containerSet,
+                service,
+            ])
+
+    },
+
 } + prompt
 

tg-launch-azure-cassandra.yaml

@@ -177,6 +177,23 @@ services:
           memory: 128M
     image: docker.io/trustgraph/trustgraph-flow:0.11.1
     restart: on-failure:100
+  metering-rag:
+    command:
+    - metering
+    - -p
+    - pulsar://pulsar:6650
+    - -i
+    - non-persistent://tg/response/text-completion-rag-response
+    deploy:
+      resources:
+        limits:
+          cpus: '0.5'
+          memory: 128M
+        reservations:
+          cpus: '0.1'
+          memory: 128M
+    image: docker.io/trustgraph/trustgraph-flow:0.11.1
+    restart: on-failure:100
   pdf-decoder:
     command:
     - pdf-decoder

tg-launch-azure-neo4j.yaml

@@ -160,6 +160,23 @@ services:
           memory: 128M
     image: docker.io/trustgraph/trustgraph-flow:0.11.1
     restart: on-failure:100
+  metering-rag:
+    command:
+    - metering
+    - -p
+    - pulsar://pulsar:6650
+    - -i
+    - non-persistent://tg/response/text-completion-rag-response
+    deploy:
+      resources:
+        limits:
+          cpus: '0.5'
+          memory: 128M
+        reservations:
+          cpus: '0.1'
+          memory: 128M
+    image: docker.io/trustgraph/trustgraph-flow:0.11.1
+    restart: on-failure:100
   neo4j:
     deploy:
       resources:

tg-launch-bedrock-cassandra.yaml

@@ -177,6 +177,23 @@ services:
           memory: 128M
     image: docker.io/trustgraph/trustgraph-flow:0.11.1
     restart: on-failure:100
+  metering-rag:
+    command:
+    - metering
+    - -p
+    - pulsar://pulsar:6650
+    - -i
+    - non-persistent://tg/response/text-completion-rag-response
+    deploy:
+      resources:
+        limits:
+          cpus: '0.5'
+          memory: 128M
+        reservations:
+          cpus: '0.1'
+          memory: 128M
+    image: docker.io/trustgraph/trustgraph-flow:0.11.1
+    restart: on-failure:100
   pdf-decoder:
     command:
     - pdf-decoder

tg-launch-bedrock-neo4j.yaml

@@ -160,6 +160,23 @@ services:
           memory: 128M
     image: docker.io/trustgraph/trustgraph-flow:0.11.1
     restart: on-failure:100
+  metering-rag:
+    command:
+    - metering
+    - -p
+    - pulsar://pulsar:6650
+    - -i
+    - non-persistent://tg/response/text-completion-rag-response
+    deploy:
+      resources:
+        limits:
+          cpus: '0.5'
+          memory: 128M
+        reservations:
+          cpus: '0.1'
+          memory: 128M
+    image: docker.io/trustgraph/trustgraph-flow:0.11.1
+    restart: on-failure:100
   neo4j:
     deploy:
       resources:

tg-launch-claude-cassandra.yaml

@@ -177,6 +177,23 @@ services:
           memory: 128M
     image: docker.io/trustgraph/trustgraph-flow:0.11.1
     restart: on-failure:100
+  metering-rag:
+    command:
+    - metering
+    - -p
+    - pulsar://pulsar:6650
+    - -i
+    - non-persistent://tg/response/text-completion-rag-response
+    deploy:
+      resources:
+        limits:
+          cpus: '0.5'
+          memory: 128M
+        reservations:
+          cpus: '0.1'
+          memory: 128M
+    image: docker.io/trustgraph/trustgraph-flow:0.11.1
+    restart: on-failure:100
   pdf-decoder:
     command:
     - pdf-decoder

tg-launch-claude-neo4j.yaml

@@ -160,6 +160,23 @@ services:
           memory: 128M
     image: docker.io/trustgraph/trustgraph-flow:0.11.1
     restart: on-failure:100
+  metering-rag:
+    command:
+    - metering
+    - -p
+    - pulsar://pulsar:6650
+    - -i
+    - non-persistent://tg/response/text-completion-rag-response
+    deploy:
+      resources:
+        limits:
+          cpus: '0.5'
+          memory: 128M
+        reservations:
+          cpus: '0.1'
+          memory: 128M
+    image: docker.io/trustgraph/trustgraph-flow:0.11.1
+    restart: on-failure:100
   neo4j:
     deploy:
       resources:

tg-launch-cohere-cassandra.yaml

@@ -177,6 +177,23 @@ services:
           memory: 128M
     image: docker.io/trustgraph/trustgraph-flow:0.11.1
     restart: on-failure:100
+  metering-rag:
+    command:
+    - metering
+    - -p
+    - pulsar://pulsar:6650
+    - -i
+    - non-persistent://tg/response/text-completion-rag-response
+    deploy:
+      resources:
+        limits:
+          cpus: '0.5'
+          memory: 128M
+        reservations:
+          cpus: '0.1'
+          memory: 128M
+    image: docker.io/trustgraph/trustgraph-flow:0.11.1
+    restart: on-failure:100
   pdf-decoder:
     command:
     - pdf-decoder

tg-launch-cohere-neo4j.yaml

@@ -160,6 +160,23 @@ services:
           memory: 128M
     image: docker.io/trustgraph/trustgraph-flow:0.11.1
     restart: on-failure:100
+  metering-rag:
+    command:
+    - metering
+    - -p
+    - pulsar://pulsar:6650
+    - -i
+    - non-persistent://tg/response/text-completion-rag-response
+    deploy:
+      resources:
+        limits:
+          cpus: '0.5'
+          memory: 128M
+        reservations:
+          cpus: '0.1'
+          memory: 128M
+    image: docker.io/trustgraph/trustgraph-flow:0.11.1
+    restart: on-failure:100
   neo4j:
     deploy:
       resources:

tg-launch-llamafile-cassandra.yaml

@@ -177,6 +177,23 @@ services:
           memory: 128M
     image: docker.io/trustgraph/trustgraph-flow:0.11.1
     restart: on-failure:100
+  metering-rag:
+    command:
+    - metering
+    - -p
+    - pulsar://pulsar:6650
+    - -i
+    - non-persistent://tg/response/text-completion-rag-response
+    deploy:
+      resources:
+        limits:
+          cpus: '0.5'
+          memory: 128M
+        reservations:
+          cpus: '0.1'
+          memory: 128M
+    image: docker.io/trustgraph/trustgraph-flow:0.11.1
+    restart: on-failure:100
   pdf-decoder:
     command:
     - pdf-decoder

tg-launch-llamafile-neo4j.yaml

@@ -160,6 +160,23 @@ services:
           memory: 128M
     image: docker.io/trustgraph/trustgraph-flow:0.11.1
     restart: on-failure:100
+  metering-rag:
+    command:
+    - metering
+    - -p
+    - pulsar://pulsar:6650
+    - -i
+    - non-persistent://tg/response/text-completion-rag-response
+    deploy:
+      resources:
+        limits:
+          cpus: '0.5'
+          memory: 128M
+        reservations:
+          cpus: '0.1'
+          memory: 128M
+    image: docker.io/trustgraph/trustgraph-flow:0.11.1
+    restart: on-failure:100
   neo4j:
     deploy:
       resources:

tg-launch-ollama-cassandra.yaml

@@ -177,6 +177,23 @@ services:
           memory: 128M
     image: docker.io/trustgraph/trustgraph-flow:0.11.1
     restart: on-failure:100
+  metering-rag:
+    command:
+    - metering
+    - -p
+    - pulsar://pulsar:6650
+    - -i
+    - non-persistent://tg/response/text-completion-rag-response
+    deploy:
+      resources:
+        limits:
+          cpus: '0.5'
+          memory: 128M
+        reservations:
+          cpus: '0.1'
+          memory: 128M
+    image: docker.io/trustgraph/trustgraph-flow:0.11.1
+    restart: on-failure:100
   pdf-decoder:
     command:
     - pdf-decoder

tg-launch-ollama-neo4j.yaml

@@ -160,6 +160,23 @@ services:
           memory: 128M
     image: docker.io/trustgraph/trustgraph-flow:0.11.1
     restart: on-failure:100
+  metering-rag:
+    command:
+    - metering
+    - -p
+    - pulsar://pulsar:6650
+    - -i
+    - non-persistent://tg/response/text-completion-rag-response
+    deploy:
+      resources:
+        limits:
+          cpus: '0.5'
+          memory: 128M
+        reservations:
+          cpus: '0.1'
+          memory: 128M
+    image: docker.io/trustgraph/trustgraph-flow:0.11.1
+    restart: on-failure:100
   neo4j:
     deploy:
       resources:

tg-launch-openai-cassandra.yaml

@@ -177,6 +177,23 @@ services:
           memory: 128M
     image: docker.io/trustgraph/trustgraph-flow:0.11.1
     restart: on-failure:100
+  metering-rag:
+    command:
+    - metering
+    - -p
+    - pulsar://pulsar:6650
+    - -i
+    - non-persistent://tg/response/text-completion-rag-response
+    deploy:
+      resources:
+        limits:
+          cpus: '0.5'
+          memory: 128M
+        reservations:
+          cpus: '0.1'
+          memory: 128M
+    image: docker.io/trustgraph/trustgraph-flow:0.11.1
+    restart: on-failure:100
   pdf-decoder:
     command:
     - pdf-decoder

tg-launch-openai-neo4j.yaml

@@ -160,6 +160,23 @@ services:
           memory: 128M
     image: docker.io/trustgraph/trustgraph-flow:0.11.1
     restart: on-failure:100
+  metering-rag:
+    command:
+    - metering
+    - -p
+    - pulsar://pulsar:6650
+    - -i
+    - non-persistent://tg/response/text-completion-rag-response
+    deploy:
+      resources:
+        limits:
+          cpus: '0.5'
+          memory: 128M
+        reservations:
+          cpus: '0.1'
+          memory: 128M
+    image: docker.io/trustgraph/trustgraph-flow:0.11.1
+    restart: on-failure:100
   neo4j:
     deploy:
       resources:

tg-launch-vertexai-cassandra.yaml

@@ -177,6 +177,23 @@ services:
           memory: 128M
     image: docker.io/trustgraph/trustgraph-flow:0.11.1
     restart: on-failure:100
+  metering-rag:
+    command:
+    - metering
+    - -p
+    - pulsar://pulsar:6650
+    - -i
+    - non-persistent://tg/response/text-completion-rag-response
+    deploy:
+      resources:
+        limits:
+          cpus: '0.5'
+          memory: 128M
+        reservations:
+          cpus: '0.1'
+          memory: 128M
+    image: docker.io/trustgraph/trustgraph-flow:0.11.1
+    restart: on-failure:100
   pdf-decoder:
     command:
     - pdf-decoder

tg-launch-vertexai-neo4j.yaml

@@ -160,6 +160,23 @@ services:
           memory: 128M
     image: docker.io/trustgraph/trustgraph-flow:0.11.1
     restart: on-failure:100
+  metering-rag:
+    command:
+    - metering
+    - -p
+    - pulsar://pulsar:6650
+    - -i
+    - non-persistent://tg/response/text-completion-rag-response
+    deploy:
+      resources:
+        limits:
+          cpus: '0.5'
+          memory: 128M
+        reservations:
+          cpus: '0.1'
+          memory: 128M
+    image: docker.io/trustgraph/trustgraph-flow:0.11.1
+    restart: on-failure:100
   neo4j:
     deploy:
       resources: