signalflow-client-python

Table of Contents

• Installation
• Run a SignalFlow computation
• TLS certificates
• License

SignalFlow is the SignalFx real-time analytics computation language. The SignalFlow API allows SignalFx users to execute real-time streaming analytics computations on the SignalFx platform. For more information, see the Splunk Observability Cloud developer documentation:

• SignalFlow Overview
• SignalFlow API Reference

The SignalFlow Python library is a client that opens a connection to SignalFx, allowing you to execute SignalFlow programs against the back end and then stream data back to the client.

The following SignalFlow program returns the current number of users in your organization:

data('sf.org.num.orguser').publish()

Installation

To install the SignalFlow Python client library, open a terminal and run the following command:

pip install signalflow-client-python

Run a SignalFlow computation

The following example allows you run a SignalFlow computation from the command line. For additional examples, see the examples directory.

1. Install the signalflow-client-python package:

   pip install signalflow-client-python
   

2. Create a .py file that includes the following content:

#!/usr/bin/env python

import argparse

from signalfx.signalflow import SignalFlowClient


def main():
    parser = argparse.ArgumentParser(
        description="SignalFx SignalFlow streaming analytics demo"
    )
    parser.add_argument(
        "--stream-endpoint",
        help="SignalFx SignalFlow stream API endpoint",
        default="https://stream.signalfx.com",
    )
    parser.add_argument("token", help="Your SignalFx API access token")
    parser.add_argument("program", help="SignalFlow program to execute")
    options = parser.parse_args()
    client = SignalFlowClient(
        token=options.token,
        endpoint=options.stream_endpoint,
    )
    try:
        # Execute the computation and iterate over the message stream
        print("Requesting computation: {0}".format(options.program))
        c = client.execute(options.program)
        print("Waiting for data...")
        for msg in c.stream():
            print(f"Message: {msg}")
    except KeyboardInterrupt:
        print("Detaching from computation...")
    finally:
        client.close()
    print("Done.")


if __name__ == "__main__":
    main()

3. Run the Python script, specifying values for the streaming endpoint (optional), the API access token, and the SignalFlow program.

   For example:

   python <file-name>.py --stream-endpoint https://stream.us0.signalfx.com <api-token> "data('sf.org.num.orguser').publish()"
   

TLS certificates

The default WebSocket transport requires an https:// or wss:// endpoint. It checks the server's certificate chain and hostname before sending your token. The client uses the CA bundle from certifi by default. Keep certifi up to date to receive changes to the trusted certificate authorities.

For a private deployment, pass the path to a PEM CA bundle with ca_certs:

client = SignalFlowClient(
    token=api_token,
    endpoint="https://stream.example.com",
    ca_certs="/path/to/ca-bundle.pem",
)

This bundle replaces the default CA bundle. Include any public CA certificates that the deployment also needs. The client requires certificate and hostname checks with either bundle, including when it reconnects. The ca_certs option also works with the Server-Sent Events transport. Custom transport factories must accept this keyword when the caller supplies it.

License

signalflow-client-python is distributed under the terms of the Apache-2.0 license.