jKool Python Streaming API
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.


TNT4PY - Python Streaming

Stream your logs, metrics, custom KPIs using jKool.streaming. Search, create dashboards & analyze your data using jKool @ https://www.jkoolcloud.com.


Apache V2.0

How to Start Streaming

Easily stream simple messages from command line

  • Get started streaming quickly by calling jKool/streaming.py from the command line.
  • Specify which protocol to use for streaming with --https or --mqtt
  • See full usage with -h option.

Example stream over https:

python streaming.py "Your message to stream" --https your-access-token

Example stream over mqtt:

python streaming.py "Your message to stream" --mqtt broker-address your-username your-password --topic python/streams

Incorporate with your python applications

  • Create a HttpHandler using your jkool-api-access-token and optional url and logging level
    • Default streaming url is https://data.jkoolcloud.com
    • Default log level is logging.INFO
  • Add this handler to your python loggers
  • Logs, metrics are automatically streamed when logging calls are made.
from jKool import streaming
import logging

logger = logging.getLogger("jKool logger")
hdlr = streaming.HttpHandler("jkool-api-access-token")

logger.error("Test log")

Streaming over MQTT

  • MqttHandler implements an MQTT client using the Eclipse Paho Python Client API. See the documentation for installation instructions.
  • Create an MqttHandler with the url of the mqtt broker to publish to.
    • Default logging level is logging.INFO
    • Default message topic is the name of the logger but can be specified with topic parameter.
    • Specify a unique client id string with client_id. If not specified, one will be generated. In this case the clean_session parameter must be True.
    • If clean_session set to True the broker will remove all information about this client when it disconnects. If False, the client is a durable client and subscription information and queued messages will be retained when the client disconnects. Defaults to True
    • Use username and password to set a username/password for broker authentication.
    • Configure network encryption and authentication options by passing a dictionary or keyword arguements for ssl options. See Paho Python Client documentation for tls_set() for valid options.
  • Add this handler to your python loggers
# dictionary with ssl options
options = {"ca_certs":"certificates/ca_certs.crt", "certfile":"path-to-certfile", "keyfile":"path-to-keyfile",
"cert_reqs":"ssl.CERT_REQUIRED", "tls_version":"ssl.PROTOCOL_TLSv1", "ciphers":"a-cipher"}

# options passed as dict
MqttHandler("broker-url", topic="tnt4py example", client_id = str(uuid.uuid4()), clean_session=False, **options)

# or specify each
MqttHandler("broker-url", ca_certs="path-to-ca-file", cert_reqs=ssl.CERT_REQUIRED)

Event Stream Decoration

Events can be decorated/enriched before streaming to jKool. Use logEvent helper method with user defined decorations. Required parameters are the logger instance, logging message, and the sourcefqn. Source fully qualified name (fqn) is a cannonical event source name with the following convention TYPE=name#TYPE=name..#TYPE=name which is read from left to right and defines enclosure relationship.

Example: APPL=PythonStreaming#SERVER=PythonServer100#NETADDR=,13.41053 interpreted as application PythonStreaming running on server PythonServer100 at network address in datacenter DC1 located in geo-location 52.52437,13.41053.


streaming.logEvent(logger, "This is an example",

Use optional parameters to decorate event streams.

sourcefqn = "APPL=PythonStreaming#SERVER=PythonServer100#NETADDR=,13.41053"
streaming.logEvent(logger, "This is an example", sourcefqn,
       time_usec=int(time.time() * 1000000), corr_id="your-correlator-id", location="Atlanta, Ga")

Correlators are used to connect/stitch multiple events into a single related activity. Any number of events are related when they share one ore more correlators.

User Defined Metrics

You can also report user defined metrics (e.g. CPU, memory, Order Amount). This is done via Snapshots and Properties in the metrics module. A Snapshot holds a collection of user define Properties, each property is name, value, type pairing.

Streaming Events with Snapshots and Custom Properties

Snapshots can be attached to an Event by adding a list of snapshots to the snapshots argument.

mySnapshot = Snapshot("Payment", category="Order")
mySnapshot.addProperty("order-no", orderNo, "string")
mySnapshot.addProperty("order-amount", orderAmount, "integer")

# snapshots argument must be a list containing one or more Snapshots
logEvent(logger, "Order Processed Succesfully", sourcefqn, snapshots=[mySnapshot])

Snapshots and Properties are automatically serialized into JSON format.