Problem

A metric to easily see that a Trigger isn't working, because the trigger's sink is failing (e.g. it always returns a 4xx/5xx or times out, DNS lookup fails, etc.).

Persona:
Which persona is this feature for?
Event Consumer

Exit Criteria

Expose Trigger's sink failures count metric.

Time Estimate (optional):
How many developer-days do you think this may take to resolve?
1

Additional context (optional)

We can expose the following metric, that counts how many times we failed to send an event to a sink:

• sink_failures_total{id=<trigger_uid>} <num_failures>

  • The id label allows correlating this metric with Kafka consumer metrics like:
    kafka_consumer_request_rate{client_id="consumer-<trigger_uid>-1",kafka_version="2.6.0",} <value>
kafka_consumer_io_wait_ratio{client_id="consumer-<trigger_uid>-1",kafka_version="2.6.0",} <value>
...