Add TrafficSplitter API

test/apis/traffic-splitter/.dockerignore

@@ -0,0 +1,8 @@
+Dockerfile
+README.md
+sample.json
+*.pyc
+*.pyo
+*.pyd
+__pycache__
+.pytest_cache

test/apis/traffic-splitter/Dockerfile

@@ -0,0 +1,18 @@
+FROM python:3.8-slim
+
+# Allow statements and log messages to immediately appear in the Knative logs
+ENV PYTHONUNBUFFERED True
+
+# Install production dependencies
+RUN pip install --no-cache-dir "uvicorn[standard]" gunicorn fastapi pydantic
+
+# Copy local code to the container image.
+COPY . /app
+WORKDIR /app/
+
+ENV PYTHONPATH=/app
+ENV CORTEX_PORT=9000
+
+# Run the web service on container startup. Here we use the gunicorn
+# webserver, with one worker process and one thread.
+CMD gunicorn -k uvicorn.workers.UvicornWorker --bind :$CORTEX_PORT main:app

test/apis/traffic-splitter/cortex_cpu.yaml

@@ -0,0 +1,63 @@
+- name: prime-generator-a
+  kind: RealtimeAPI
+  pod:
+    port: 9000
+    containers:
+    - name: api
+      image: 499593605069.dkr.ecr.us-west-2.amazonaws.com/sample/realtime-prime-generator:latest
+      readiness_probe:
+        http_get:
+          path: "/healthz"
+          port: 9000
+      compute:
+        cpu: 200m
+        mem: 128M
+  autoscaling:
+    max_concurrency: 1
+
+- name: prime-generator-b
+  kind: RealtimeAPI
+  pod:
+    port: 9000
+    containers:
+    - name: api
+      image: 499593605069.dkr.ecr.us-west-2.amazonaws.com/sample/realtime-prime-generator:latest
+      readiness_probe:
+        http_get:
+          path: "/healthz"
+          port: 9000
+      compute:
+        cpu: 200m
+        mem: 128M
+  autoscaling:
+    max_concurrency: 1
+
+- name: request-recorder
+  kind: RealtimeAPI
+  pod:
+    port: 9000
+    containers:
+    - name: api
+      image: 499593605069.dkr.ecr.us-west-2.amazonaws.com/sample/realtime-prime-generator:latest
+      env:
+        RECORD_ONLY: "true"
+      readiness_probe:
+        http_get:
+          path: "/healthz"
+          port: 9000
+      compute:
+        cpu: 200m
+        mem: 128M
+  autoscaling:
+    max_concurrency: 1
+
+- name: iris-classifier
+  kind: TrafficSplitter
+  apis:
+    - name: prime-generator-a
+      weight: 30
+    - name: prime-generator-b
+      weight: 70
+    - name: request-recorder
+      shadow: true
+      weight: 100

test/apis/traffic-splitter/main.py

@@ -0,0 +1,42 @@
+import os
+from typing import DefaultDict
+
+from fastapi import FastAPI
+from pydantic import BaseModel
+
+
+def generate_primes(limit=None):
+    """Sieve of Eratosthenes"""
+    not_prime = DefaultDict(list)
+    num = 2
+    while limit is None or num <= limit:
+        if num in not_prime:
+            for prime in not_prime[num]:
+                not_prime[prime + num].append(prime)
+            del not_prime[num]
+        else:
+            yield num
+            not_prime[num * num] = [num]
+        num += 1
+
+
+class Request(BaseModel):
+    primes_to_generate: float
+
+
+app = FastAPI()
+record_only = os.getenv("RECORD_ONLY", "false")
+
+
+@app.get("/healthz")
+def healthz():
+    return "ok"
+
+
+@app.post("/")
+def prime_numbers(request: Request):
+    if record_only == "true":
+        print(f"received payload", request.primes_to_generate)
+        return
+
+    return {"prime_numbers": list(generate_primes(request.primes_to_generate))}