feat: Add FastAPI skill classification service with Prometheus monitoring and MLflow integration, updating dependencies and Docker Compose.

docker-compose.yml

@@ -47,6 +47,17 @@ services:
         condition: service_healthy
     restart: unless-stopped
 
+  prometheus:
+    image: prom/prometheus:latest
+    container_name: prometheus
+    volumes:
+      - ./monitoring/prometheus/prometheus.yml:/etc/prometheus/prometheus.yml
+    ports:
+      - "9090:9090"
+    networks:
+      - hopcroft-net
+    restart: unless-stopped
+
 networks:
   hopcroft-net:
     driver: bridge

hopcroft_skill_classification_tool_competition/main.py

@@ -22,9 +22,17 @@ import os
 import time
 from typing import List
 
-from fastapi import FastAPI, HTTPException, status
+from fastapi import FastAPI, HTTPException, status, Request, Response
 from fastapi.responses import JSONResponse, RedirectResponse
 import mlflow
+from prometheus_client import (
+    CONTENT_TYPE_LATEST,
+    Counter,
+    Gauge,
+    Histogram,
+    Summary,
+    generate_latest,
+)
 from pydantic import ValidationError
 
 from hopcroft_skill_classification_tool_competition.api_models import (
@@ -40,6 +48,34 @@ from hopcroft_skill_classification_tool_competition.api_models import (
 from hopcroft_skill_classification_tool_competition.config import MLFLOW_CONFIG
 from hopcroft_skill_classification_tool_competition.modeling.predict import SkillPredictor
 
+# Define Prometheus Metrics
+# Counter: Total number of requests
+REQUESTS_TOTAL = Counter(
+    "hopcroft_requests_total",
+    "Total number of requests",
+    ["method", "endpoint", "http_status"],
+)
+
+# Histogram: Request duration
+REQUEST_DURATION_SECONDS = Histogram(
+    "hopcroft_request_duration_seconds",
+    "Request duration in seconds",
+    ["method", "endpoint"],
+)
+
+# Gauge: In-progress requests
+IN_PROGRESS_REQUESTS = Gauge(
+    "hopcroft_in_progress_requests",
+    "Number of requests currently in progress",
+    ["method", "endpoint"],
+)
+
+# Summary: Model prediction time
+MODEL_PREDICTION_SECONDS = Summary(
+    "hopcroft_prediction_processing_seconds",
+    "Time spent processing model predictions",
+)
+
 predictor = None
 model_version = "1.0.0"
 
@@ -85,6 +121,43 @@ app = FastAPI(
 )
 
 
+@app.middleware("http")
+async def monitor_requests(request: Request, call_next):
+    """Middleware to collect Prometheus metrics for each request."""
+    method = request.method
+    # Use a simplified path or template if possible to avoid high cardinality
+    # For now, using request.url.path is acceptable for this scale
+    endpoint = request.url.path
+
+    IN_PROGRESS_REQUESTS.labels(method=method, endpoint=endpoint).inc()
+    start_time = time.time()
+
+    try:
+        response = await call_next(request)
+        status_code = response.status_code
+        REQUESTS_TOTAL.labels(
+            method=method, endpoint=endpoint, http_status=status_code
+        ).inc()
+        return response
+    except Exception as e:
+        REQUESTS_TOTAL.labels(
+            method=method, endpoint=endpoint, http_status=500
+        ).inc()
+        raise e
+    finally:
+        duration = time.time() - start_time
+        REQUEST_DURATION_SECONDS.labels(method=method, endpoint=endpoint).observe(
+            duration
+        )
+        IN_PROGRESS_REQUESTS.labels(method=method, endpoint=endpoint).dec()
+
+
+@app.get("/metrics", tags=["Observability"])
+async def metrics():
+    """Expose Prometheus metrics."""
+    return Response(content=generate_latest(), media_type=CONTENT_TYPE_LATEST)
+
+
 @app.get("/", tags=["Root"])
 async def root():
     """Return basic API information."""
@@ -143,9 +216,11 @@ async def predict_skills(issue: IssueInput) -> PredictionRecord:
 
         # Combine text fields if needed, or just use issue_text
         # The predictor expects a single string
+        # The predictor expects a single string
         full_text = f"{issue.issue_text} {issue.issue_description or ''} {issue.repo_name or ''}"
 
-        predictions_data = predictor.predict(full_text)
+        with MODEL_PREDICTION_SECONDS.time():
+            predictions_data = predictor.predict(full_text)
 
         # Convert to Pydantic models
         predictions = [

monitoring/README.md

@@ -0,0 +1,36 @@
+# Metrics Collection & Verification
+
+This directory contains the configuration for Prometheus monitoring.
+
+## Configuration
+- **Prometheus Config**: `prometheus/prometheus.yml`
+- **Scrape Target**: `hopcroft-api:8080`
+- **Metrics Endpoint**: `http://localhost:8080/metrics`
+
+## Verification Queries (PromQL)
+
+You can run these queries in the Prometheus Expression Browser (`http://localhost:9090/graph`):
+
+### 1. Request Rate (Counter)
+Shows the rate of requests per second over the last minute.
+```promql
+rate(hopcroft_requests_total[1m])
+```
+
+### 2. Average Request Duration (Histogram)
+Calculates average latency.
+```promql
+rate(hopcroft_request_duration_seconds_sum[5m]) / rate(hopcroft_request_duration_seconds_count[5m])
+```
+
+### 3. Current In-Progress Requests (Gauge)
+Shows how many requests are currently being processed.
+```promql
+hopcroft_in_progress_requests
+```
+
+### 4. Model Prediction Time (Summary)
+Shows the 90th percentile of model prediction time.
+```promql
+hopcroft_prediction_processing_seconds{quantile="0.9"}
+```

monitoring/prometheus/prometheus.yml

@@ -0,0 +1,7 @@
+global:
+  scrape_interval: 15s
+
+scrape_configs:
+  - job_name: 'hopcroft-api'
+    static_configs:
+      - targets: ['hopcroft-api:8080']

requirements.txt

@@ -23,6 +23,7 @@ sentence-transformers
 
 # API Framework
 fastapi[standard]>=0.115.0
+prometheus-client>=0.17.0
 pydantic>=2.0.0
 uvicorn>=0.30.0
 httpx>=0.27.0