Python library for building InfraSonar Probes

This library is created for building InfraSonar probes.

Environment variable

Usage

Building an InfraSonar.get_state

import logging
from libprobe.asset import Asset
from libprobe.probe import Probe
from libprobe.severity import Severity
from libprobe.exceptions import (
    CheckException,
    IgnoreResultException,
    IgnoreCheckException,
    IncompleteResultException,
    NoCountException,
)

__version__ = "0.1.0"


async def my_first_check(asset: Asset, asset_config: dict, check_config: dict):
    """My first check.
    Arguments:
      asset:        Asset contains an id, name and check which should be used
                    for logging;
      asset_config: local configuration for this asset, for example credentials;
      check_config: configuration for this check; contains for example the
                    interval at which the check is running and an address of
                    the asset to probe;
    """
    if "ignore_this_check_iteration":
        # nothing will be send to InfraSonar for this check iteration;
        raise IgnoreResultException()

    if "no_longer_try_this_check":
        # nothing will be send to InfraSonar for this check iteration and the
        # check will not start again until the probe restarts or configuration
        # has been changed;
        raise IgnoreCheckException()

    if "something_has_happened":
        # send a check error to InfraSonar because something has happened which
        # prevents us from building a check result; The default severity for a
        # CheckException is MEDIUM but this can be overwritten;
        raise CheckException("something went wrong", severity=Severity.LOW)

    if "something_unexpected_has_happened":
        # other exceptions will be converted to CheckException, MEDIUM severity
        raise Exception("something went wrong")

    # A check result may have multiple types, items, and/or metrics
    result = {"myType": [{"name": "my item"}]}

    if "result_is_incomplete":
        # optionally, IncompleteResultException can be given another severity;
        # the default severity is LOW.
        raise IncompleteResultException('missing type x', result)

    if "not_count_as_check_result":
        # optionally, NoCountException can be raised in which case the check
        # result is not counted by InfraSonar; Thus, the last seen services
        # will not "see" this check result.
        # A severity can be given if we also want a check error; (similar to
        # the IncompleteResultException exception)
        raise NoCountException('do not count this check result', result)

    # Use the asset in logging; this will include asset info and the check key
    logging.info(f"log something; {asset}")

    # Return the check result
    return result


if __name__ == "__main__":
    checks = {
        "myFirstCheck": my_first_check,
    }

    # Initialize the probe with a name, version and checks
    probe = Probe("myProbe", __version__, checks)

    # Start the probe
    probe.start()

ASCII item names

InfraSonar requires each item to have a unique name property. The value for name must be a string with ASCII compatible character. When your name is not guaranteed to be ASCII compatible, the following code replaces the incompatible characters with question marks (?):

name = name.encode('ascii', errors='replace').decode()

Config

When using a password or secret within a config section, the library will encrypt the value so it will be unreadable by users. This must not be regarded as true encryption as the encryption key is publicly available.

Example yaml configuration:

exampleProbe:
  config:
    username: alice
    password: secret_password
  assets:
  - id: 123
    config:
      username: bob
      password: "my secret"
  - id: [456, 789]
    config:
      username: charlie
      password: "my other secret"
otherProbe:
  use: exampleProbe  # use the exampleProbe config for this probe

Dry run

Create a yaml file, for example (test.yaml):

asset:
  name: "foo.local"
  check: "system"
  config:
    address: "192.168.1.2"

Run the probe with the DRY_RUN environment variable set the the yaml file above.

DRY_RUN=test.yaml python main.py

Note: Optionally an asset id might be given which can by used to find asset configuration in the local asset configuration file. Asset config is also optional.

Dump to JSON

A dry run writes all log to stderr and only the JSON dump is written to stdout. Therefore, writing the output to JSON is easy:

DRY_RUN=test.yaml python main.py > dump.json