Observability

The app bridges two SDK hooks onto Django signals: poli_page_retry fires before each retry, and poli_page_error fires when a request fails for good. Connect receivers to them to log, count or trace. They are the integration’s observability hooks.

Signals

Both signals live in django_poli_page.signals. The sender is the django_poli_page.apps.PoliPageConfig class, not an instance.

• poli_page_retry is sent before each SDK retry sleep. Receivers get event, a poli_page.RetryEvent with attempt (the attempt about to be made, 2 for the first retry), delay_ms (the sleep before it) and reason (the PoliPageError that caused the retry).
• poli_page_error is sent on terminal failure, when the retries are exhausted or the error is not retryable. Receivers get error, the PoliPageError.

The SDK does not wait for a receiver, and a receiver cannot change the request.

Use your own callables

ON_RETRY and ON_ERROR in POLI_PAGE take a dotted path to a callable. When one is set, your callable replaces the signal for that hook: it receives the RetryEvent or the PoliPageError directly, and the signal is not sent. See Configuration.

The app does not expose the SDK’s on_request and on_response hooks.

Request IDs

error.request_id is the API’s x-request-id. Log it: it is what Poli Page support asks for. The error handler also sends it as requestId in the JSON body of every error answer.

Example

invoices/signals.py

import logging

from django.dispatch import receiver

from django_poli_page.signals import poli_page_error, poli_page_retry

logger = logging.getLogger("poli_page")


@receiver(poli_page_retry)
def log_retry(sender, event, **kwargs):
    logger.warning("retry %s in %sms: %s", event.attempt, event.delay_ms, event.reason.code)


@receiver(poli_page_error)
def log_error(sender, error, **kwargs):
    logger.error("%s request_id=%s: %s", error.code, error.request_id or "-", error.message)

Import the module from an AppConfig.ready() so the receivers are connected at startup.

See also

• Configuration
• Errors