airflow.providers.datadog.hooks.datadog

Module Contents

class airflow.providers.datadog.hooks.datadog.DatadogHook(datadog_conn_id: str = 'datadog_default')[source]

Bases: airflow.hooks.base.BaseHook, airflow.utils.log.logging_mixin.LoggingMixin

Uses datadog API to send metrics of practically anything measurable, so it's possible to track # of db records inserted/deleted, records read from file and many other useful metrics.

Depends on the datadog API, which has to be deployed on the same server where Airflow runs.

Parameters

datadog_conn_id (str) -- The connection to datadog, containing metadata for api keys.

validate_response(self, response: Dict[str, Any])[source]

Validate Datadog response

send_metric(self, metric_name: str, datapoint: Union[float, int], tags: Optional[List[str]] = None, type_: Optional[str] = None, interval: Optional[int] = None)[source]

Sends a single datapoint metric to DataDog

Parameters

  • metric_name (str) -- The name of the metric

  • datapoint (int or float) -- A single integer or float related to the metric

  • tags (list) -- A list of tags associated with the metric

  • type (str) -- Type of your metric: gauge, rate, or count

  • interval (int) -- If the type of the metric is rate or count, define the corresponding interval

query_metric(self, query: str, from_seconds_ago: int, to_seconds_ago: int)[source]

Queries datadog for a specific metric, potentially with some function applied to it and returns the results.

Parameters

  • query (str) -- The datadog query to execute (see datadog docs)

  • from_seconds_ago (int) -- How many seconds ago to start querying for.

  • to_seconds_ago (int) -- Up to how many seconds ago to query for.

post_event(self, title: str, text: str, aggregation_key: Optional[str] = None, alert_type: Optional[str] = None, date_happened: Optional[int] = None, handle: Optional[str] = None, priority: Optional[str] = None, related_event_id: Optional[int] = None, tags: Optional[List[str]] = None, device_name: Optional[List[str]] = None)[source]

Posts an event to datadog (processing finished, potentially alerts, other issues) Think about this as a means to maintain persistence of alerts, rather than alerting itself.

Parameters

  • title (str) -- The title of the event

  • text (str) -- The body of the event (more information)

  • aggregation_key (str) -- Key that can be used to aggregate this event in a stream

  • alert_type (str) -- The alert type for the event, one of ["error", "warning", "info", "success"]

  • date_happened (int) -- POSIX timestamp of the event; defaults to now

  • handle -- str

  • priority (str) -- Priority to post the event as. ("normal" or "low", defaults to "normal")

  • related_event_id (id) -- Post event as a child of the given event

  • tags (list[str]) -- List of tags to apply to the event

  • device_name (list) -- device_name to post the event with

Handle

User to post the event as; defaults to owner of the application key used to submit.

Was this entry helpful?