Web front-end and framework for monitoring an icecream distributed build cluster. See icecc/icecream.

Basic usage:

python monitor.py <scheduler_hostname> <scheduler_port>

Then open visage.html in your browser.

websocket-server - pip install websocket-server

Icemon API: All messages implemented as protocol version 22. Protocol version negotiation not supported.

Generic monitor: Fully implemented with generic publisher infrastructure.

Websocket Publisher: Implemented to send a json representation of the cluster down any connected websockets.

Web front-end: Not currently hosted by the monitor/publisher itself. But otherwise fundamentally works. Displays the cluster activity using d3.js force graph.

Message (and subclasses): Represent a single message from the scheduler. They each know how to decode themselves from wire-data.

Usage:

from pyicemon import messages

# Unpack a message from a byte string. (string must contain only exactly one message)
m = messages.unpack(wire_data)

# Pack message to a byte string.
data = m.pack()

Connection: Handles a persistent connection to a scheduler. Reads messages off the wire and returns them as parsed Message types.

Usage:

import pyicemon
from pyicemon import messages

scheduler_host, scheduler_port = "scheduler.com", 8765

# Open a connection to a server. (Also handles version negotiation)
conn = pyicemon.Connection(scheduler_host, scheduler_port)

# Send a message.
m = messages.LoginMessage()
conn.send_message(m)

# Receive messages.
while True:
    m = conn.get_message()
    print m

Monitor: Maintains a view of the current state of the cluster by reading messages from a Connection and updating itself accordingly. Reports changes in cluster state to a Publisher.

Usage:

import pyicemon
from pyicemon.publishers import WebsocketPublisher

scheduler_host, scheduler_port = "scheduler.com", 8765
mon = pyicemon.Monitor(scheduler_host, scheduler_port)
mon.addPublisher(WebsocketPublisher(port=9999))
mon.run() # Blocks forever.

Publisher: Publishes information about the cluster to an outside source.

• Implement protocol negotiation and interoperability between different protocol versions.
• Generally improve quality of web view.
• Automatic scheduler discovery.