Skip to content

Integrations

ArqonHPO integrates with your existing infrastructure.

Monitoring & Observability

Integration Status Description
Prometheus ✅ Built-in Metrics via --metrics-addr
Grafana ✅ Compatible Use Prometheus data source
OpenTelemetry 🔜 Planned v0.4 roadmap

→ Observability Guide

ML Frameworks

Integration Status Description
Ray Tune ✅ Custom Searcher Ray Guide
MLflow 🔜 Planned Tracking plugin
Weights & Biases 🔜 Planned Callback

Web Frameworks

Integration Status Description
FastAPI ✅ Example FastAPI Guide
Flask ✅ Compatible Similar to FastAPI
Django ✅ Compatible Use management commands

Infrastructure

Integration Status Description
Kubernetes ✅ Patterns K8s Guide
Docker ✅ Example See K8s guide
Helm 🔜 Planned Official chart in v0.4

Databases

ArqonHPO stores state in JSON files by default. For persistent storage:

Integration Status Description
Redis ✅ Use --state Store state in Redis via wrapper
PostgreSQL ✅ Use artifacts Import/export via SQL
S3 ✅ Use artifacts Store artifacts in S3

Example: S3 State Storage

import boto3
import json
from arqonhpo import ArqonSolver

s3 = boto3.client('s3')

def load_state(bucket, key):
    resp = s3.get_object(Bucket=bucket, Key=key)
    return resp['Body'].read().decode()

def save_state(bucket, key, state):
    s3.put_object(Bucket=bucket, Key=key, Body=state)

# Load existing or create new
try:
    state = load_state('my-bucket', 'arqon/state.json')
    solver = ArqonSolver.from_state(state)
except:
    solver = ArqonSolver(json.dumps(config))

# ... optimization loop ...

# Save state
save_state('my-bucket', 'arqon/state.json', solver.export())

Message Queues

Integration Status Description
RabbitMQ ✅ JSONL Use interactive mode
Kafka ✅ JSONL Use interactive mode
Redis Streams ✅ JSONL Use interactive mode

Example: RabbitMQ

import pika
import json
import subprocess

connection = pika.BlockingConnection()
channel = connection.channel()

channel.queue_declare(queue='arqon_ask')
channel.queue_declare(queue='arqon_results')

# Start interactive mode
proc = subprocess.Popen(
    ['arqonhpo', 'interactive', '--config', 'config.json'],
    stdin=subprocess.PIPE,
    stdout=subprocess.PIPE
)

def on_result(ch, method, props, body):
    result = json.loads(body)
    proc.stdin.write(json.dumps({"cmd": "tell", "results": [result]}).encode() + b'\n')
    proc.stdin.flush()

    # Ask for next
    proc.stdin.write(json.dumps({"cmd": "ask", "batch": 1}).encode() + b'\n')
    proc.stdin.flush()
    response = proc.stdout.readline()

    channel.basic_publish(exchange='', routing_key='arqon_ask', body=response)

channel.basic_consume(queue='arqon_results', on_message_callback=on_result)
channel.start_consuming()

CI/CD

Integration Status Description
GitHub Actions ✅ CLI Run optimization in workflows
GitLab CI ✅ CLI Same as GitHub
ArqonShip ✅ Built-in Self-healing CI

→ ArqonShip Docs

Next Steps