otupy.apps.ctxd

The discovery application recursively queries context actuators to build a chain of interconnected services. It starts from one or more root services (which consumers are known a priori) and periodically and iteratively queries additional peers found in the responses.

The application is designed to publish context data to multiple sinks.

Setup

To run the discovery, either download the source code or install from PyPI (see setup).

Configuration

The discovery application works according to a yaml configuration file that contains the following parameters:

• name: An identifier used to distinguish context originated by different ``discovery``s

• frequency: The time interval before starting a new round of queries (the real interval will be longer because the timer starts after receving the last answer). Run one-shot if sets to 0.

• loop: Number of times to repeat the discovery. Loops forever if set to -1, does not run if set to 0.

• publishers: a dictionary of places where the context data are published after each round. The configuration changes according to the specific publisher:

  • mongodb: A dictionary with configuration to write data to MongoDB

    • host: IP address or hostname of the server hosting the database

    • port: Port number where the mongodb service listens to

    • db_name: Name of the internal database to store data

    • collection: Name to be used for the collection of documents

    • user: username to connect to the database

    • pass: password to connect to the database

  • kafka: A dictionary with the configuration to publish to Kafka brokers

    • bootstrap: a list of bootstrap servers, given as:

      • host: IP address or hostname of the server hosting one bootstrap server

      • port: Port number where the bootstrap server listens to

    • topic: The topic used to publish data

    • security_protocol: Security protocol used to connect to Kafka (PLAINTEXT, SSL, SASL_PLAINTEXT, SASL_SSL - see Kafka documentation)

    • sasl_mechanism: Username/password authentication mechanism (PLAIN, GSSAPI, OAUTHBEARER, SCRAM-SHA-256, SCRAM-SHA-512). Only valid for security protocols SASL_PLAINTEXT or SASL_SSL

    • sasl_plain_username: Username when sasl mechanism is enabled

    • sasl_plain_password: Password when sasl mechanism is enabled

    • ssl_cafile: CA file used to sign the Kafka certificate

    • ssl_check_hostname: Enable (True) or disable (False) server name validation in the certificate file

  • file: A dictionary with configuration to write to file (append mode)

    • name: Filename (created if does not exist)

    • path: Filesystem path to the file (current directory if not given)

• services: A list of “root services” to query, indicated as their Consumer endpoints:

  • host

  • port

  • profile

  • encoding

  • transfer

  • endpoint

  • actuator (x-ctxd Specifiers)

• logger: The configuration for the Logging framework. See the module documentation

A template configuration file is available here.

Run

Run the discovery service:

python3 discovery.py [-c | --config <config.yaml>] [ -p | --port <port>] [ --host <hostname> ] [  --api ]

Enable the API service with the ``--api`` flag. If the hostname/port are not given, default values will be used.
Even if the API service expects a configuration file in the start command, the configuration file remains a valid option and
can be used to load "default" values and parameters that makes no sense in the API service (e.g., the ``logger`` configuration).

API service

The following endpoints are available:

• /start: POST request with a config in json format to start discovery, returns the id of started thread

• /stop: POST request with a list of thread ids to stop

• /threads: GET request which returns the list of active threads (active means not cancelled, they may have terminated their job)

• /clean: POST request that cleans up all threads

All POST requests MUST have the Content-type header set to application/json. The configuration of the start request only includes a subset of the parameters described above: logger is not allowed and is ignored.

defaults	Manage default values
graphics	Graphics tools
publishers	Utilities to export context