feat(quickstart): use unified meters (#652)

openmeterio · Mar 4, 2024 · 9c33804 · 9c33804
1 parent fc3a69e
commit 9c33804
Show file tree

Hide file tree

Showing 7 changed files with 62 additions and 50 deletions.
diff --git a/config.example.yaml b/config.example.yaml
@@ -18,10 +18,21 @@ telemetry:
 #   tokenSecret: this-isnt-secure
 
 meters:
-  - slug: m1
-    eventType: api-calls
-    valueProperty: $.duration_ms
+  # Sample meter to count API requests
+  - slug: api_requests_total        # Unique identifier for the meter
+    description: API Requests
+    eventType: request              # Filter events by type
+    aggregation: COUNT              # Aggregation method: COUNT, SUM, etc.
+    groupBy:
+      method: $.method              # HTTP Method: GET, POST, etc.
+      route: $.route                # Route: /products/:product_id
+
+  # Sample meter to count LLM Token Usage
+  - slug: tokens_total
+    description: AI Token Usage
+    eventType: prompt               # Filter events by type
     aggregation: SUM
+    valueProperty: $.tokens         # JSONPath to parse usage value
     groupBy:
-      method: $.method
-      path: $.path
+      model: $.model                # AI model used: gpt4-turbo, etc.
+      type: $.type                  # Prompt type: input, output, system
diff --git a/examples/ingest-logs/README.md b/examples/ingest-logs/README.md
@@ -18,6 +18,6 @@ Transformations in Vector uses the [Vector Remap Language](https://vector.dev/do
 
 1. First, start OpenMeter. Refer to the [quickstart guide](/quickstart) for instructions.
 2. Execute `docker compose up` within this example directory.
-3. To query meters, use the following command: `curl http://localhost:8888/api/v1/meters/m1/values`.
+3. To query meters, use the following command: `curl http://localhost:8888/api/v1/meters/api_requests_total/values`.
 
 Note: It's important that you run quickstart's `docker compose` first.
diff --git a/quickstart/README.md b/quickstart/README.md
@@ -31,15 +31,14 @@ curl -X POST http://localhost:8888/api/v1/events \
 --data-raw '
 {
   "specversion" : "1.0",
-  "type": "api-calls",
+  "type": "request",
   "id": "00001",
   "time": "2023-01-01T00:00:00.001Z",
   "source": "service-0",
   "subject": "customer-1",
   "data": {
-    "duration_ms": "1",
     "method": "GET",
-    "path": "/hello"
+    "route": "/hello"
   }
 }
 '
@@ -53,15 +52,14 @@ curl -X POST http://localhost:8888/api/v1/events \
 --data-raw '
 {
   "specversion" : "1.0",
-  "type": "api-calls",
+  "type": "request",
   "id": "00002",
   "time": "2023-01-01T00:00:00.001Z",
   "source": "service-0",
   "subject": "customer-1",
   "data": {
-    "duration_ms": "1",
     "method": "GET",
-    "path": "/hello"
+    "route": "/hello"
   }
 }
 '
@@ -75,15 +73,14 @@ curl -X POST http://localhost:8888/api/v1/events \
 --data-raw '
 {
   "specversion" : "1.0",
-  "type": "api-calls",
+  "type": "request",
   "id": "00003",
   "time": "2023-01-02T00:00:00.001Z",
   "source": "service-0",
   "subject": "customer-1",
   "data": {
-    "duration_ms": "1",
     "method": "GET",
-    "path": "/hello"
+    "route": "/hello"
   }
 }
 '
@@ -94,7 +91,7 @@ curl -X POST http://localhost:8888/api/v1/events \
 Query the usage hourly:
 
 ```sh
-curl 'http://localhost:8888/api/v1/meters/m1/query?windowSize=HOUR&groupBy=method&groupBy=path' | jq
+curl 'http://localhost:8888/api/v1/meters/api_requests_total/query?windowSize=HOUR&groupBy=method&groupBy=route' | jq
 ```
 
 ```json
@@ -108,7 +105,7 @@ curl 'http://localhost:8888/api/v1/meters/m1/query?windowSize=HOUR&groupBy=metho
       "subject": null,
       "groupBy": {
         "method": "GET",
-        "path": "/hello"
+        "route": "/hello"
       }
     },
     {
@@ -118,7 +115,7 @@ curl 'http://localhost:8888/api/v1/meters/m1/query?windowSize=HOUR&groupBy=metho
       "subject": null,
       "groupBy": {
         "method": "GET",
-        "path": "/hello"
+        "route": "/hello"
       }
     }
   ]
@@ -128,7 +125,7 @@ curl 'http://localhost:8888/api/v1/meters/m1/query?windowSize=HOUR&groupBy=metho
 Query the total usage for `customer-1`:
 
 ```sh
-curl 'http://localhost:8888/api/v1/meters/m1/query?subject=customer-1' | jq
+curl 'http://localhost:8888/api/v1/meters/api_requests_total/query?subject=customer-1' | jq
 ```
 
 ```json
@@ -147,21 +144,25 @@ curl 'http://localhost:8888/api/v1/meters/m1/query?subject=customer-1' | jq
 
 ## 4. Configure additional meter(s) _(optional)_
 
-Configure how OpenMeter should process your usage events.
-In this example we will meter the execution duration per API invocation, groupped by method and path.
-You can think about it how AWS Lambda [charges](https://aws.amazon.com/lambda/pricing/) by execution duration on a millisecond level.
+In this example we will meter LLM token usage, groupped by AI model and prompt type.
+You can think about it how OpenAI [charges](https://openai.com/pricing) by tokens for ChatGPT.
+
+Configure how OpenMeter should process your usage events in this new `tokens_total` meter.
 
 ```yaml
 # ...
 
 meters:
-  - slug: m1
-    eventType: api-calls
-    valueProperty: $.duration_ms
+  # Sample meter to count LLM Token Usage
+  - slug: tokens_total
+    description: AI Token Usage
+    eventType: prompt               # Filter events by type
     aggregation: SUM
+    valueProperty: $.tokens         # JSONPath to parse usage value
     groupBy:
-      method: $.method
-      path: $.path
+      model: $.model                # AI model used: gpt4-turbo, etc.
+      type: $.type                  # Prompt type: input, output, system
+
 ```
 
 ## Cleanup

diff --git a/quickstart/config.yaml b/quickstart/config.yaml
@@ -11,21 +11,24 @@ sink:
   namespaceRefetch: 1s
 
 meters:
-  - slug: m1
-    description: API calls
-    eventType: api-calls
-    valueProperty: $.duration_ms
-    aggregation: SUM
+  # Sample meter to count API requests
+  - slug: api_requests_total         # Unique identifier for the meter
+    description: API Requests
+    eventType: request              # Filter events by type
+    aggregation: COUNT              # Aggregation method: COUNT, SUM, etc.
     groupBy:
-      method: $.method
-      path: $.path
-  - slug: m2
-    description: Open AI Tokens
-    eventType: openai
-    valueProperty: $.total_tokens
+      method: $.method              # HTTP Method: GET, POST, etc.
+      route: $.route                # Route: /products/:product_id
+
+  # Sample meter to count LLM Token Usage
+  - slug: tokens_total
+    description: AI Token Usage
+    eventType: prompt               # Filter events by type
     aggregation: SUM
+    valueProperty: $.tokens         # JSONPath to parse usage value
     groupBy:
-      model: $.model
+      model: $.model                # AI model used: gpt4-turbo, etc.
+      type: $.type                  # Prompt type: input, output, system
 
 portal:
   enabled: true

diff --git a/quickstart/docker-compose.yaml b/quickstart/docker-compose.yaml
@@ -16,7 +16,7 @@ services:
     volumes:
       - ./config.yaml:/etc/openmeter/config.yaml
     healthcheck:
-      test: wget --no-verbose --tries=1 --spider http://openmeter:8888/api/v1/meters/m1/query || exit 1
+      test: wget --no-verbose --tries=1 --spider http://openmeter:8888/api/v1/meters/api_requests_total/query || exit 1
       interval: 5s
       timeout: 3s
       retries: 100

diff --git a/quickstart/ingest.sh b/quickstart/ingest.sh
@@ -5,15 +5,14 @@ curl -X POST http://localhost:8888/api/v1/events \
     --data-raw '
 {
   "specversion" : "1.0",
-  "type": "api-calls",
+  "type": "request",
   "id": "00001",
   "time": "2023-01-01T00:00:00.001Z",
   "source": "service-0",
   "subject": "customer-1",
   "data": {
-    "duration_ms": "1",
     "method": "GET",
-    "path": "/hello"
+    "route": "/hello"
   }
 }
 '
@@ -23,15 +22,14 @@ curl -X POST http://localhost:8888/api/v1/events \
     --data-raw '
 {
   "specversion" : "1.0",
-  "type": "api-calls",
+  "type": "request",
   "id": "00002",
   "time": "2023-01-01T00:00:00.001Z",
   "source": "service-0",
   "subject": "customer-1",
   "data": {
-    "duration_ms": "1",
     "method": "GET",
-    "path": "/hello"
+    "route": "/hello"
   }
 }
 '
@@ -41,15 +39,14 @@ curl -X POST http://localhost:8888/api/v1/events \
     --data-raw '
 {
   "specversion" : "1.0",
-  "type": "api-calls",
+  "type": "request",
   "id": "00003",
   "time": "2023-01-02T00:00:00.001Z",
   "source": "service-0",
   "subject": "customer-1",
   "data": {
-    "duration_ms": "1",
     "method": "GET",
-    "path": "/hello"
+    "route": "/hello"
   }
 }
 '
diff --git a/quickstart/verify.sh b/quickstart/verify.sh
@@ -2,7 +2,7 @@
 
 temp_file=$(mktemp)
 
-curl http://localhost:8888/api/v1/meters/m1/query?windowSize=HOUR >$temp_file
+curl http://localhost:8888/api/v1/meters/api_requests_total/query?windowSize=HOUR >$temp_file
 
 [ $(cat $temp_file | jq '.data[0].value') == 2 ] || (echo "Unexpected value" && cat $temp_file && exit 1)
 [ $(cat $temp_file | jq '.data[1].value') == 1 ] || (echo "Unexpected value" && cat $temp_file && exit 1)