Google Pub/Sub

Información general

Google Cloud Pub/Sub lleva a la nube la escalabilidad, la flexibilidad y la fiabilidad del middleware empresarial dirigido a mensajes.

Obtén métricas de Google Pub/Sub para:

  • Visualizar el rendimiento de tus temas y suscripciones Pub/Sub.
  • Correlacionar el rendimiento de tus temas y suscripciones Pub/Sub con tus aplicaciones.

Configuración

Recopilación de métricas

Instalación

Si aún no lo has hecho, primero configura la integración Google Cloud Platform. No es necesario realizar ningún otro paso de instalación.

Configuración

Para recopilar etiquetas (labels) Pub/Sub personalizadas como etiquetas (tags), activa el permiso de inventario de recursos en la nube.

APM

Los logs de Google Cloud Pub/Sub se recopilan con Google Cloud Logging y se envían a una tarea de Dataflow a través de un tema Cloud Pub/Sub. Si aún no lo has hecho, configura la generación de logs con la plantilla Dataflow de Datadog.

Una vez hecho esto, exporta tus logs de Google Cloud Pub/Sub de Google Cloud Logging al tema Pub/Sub:

  1. Ve a la página de Google Cloud Logging y filtra logs de Google Cloud Pub/Sub.

  2. Haz clic en Create sink (Crear sumidero) y asigna al sumidero el nombre correspondiente.

  3. Elige “Cloud Pub/Sub” como destino y selecciona el tema Pub/Sub creado para tal fin. Nota: El tema Pub/Sub puede encontrarse en un proyecto diferente.

    {< img src=“integrations/google_cloud_pubsub/creating_sink2.png” alt=“Exportar logs de Google Cloud Pub/Sub a Pub Sub” >}}

  4. Haz clic en Create (Crear) y espera a que aparezca el mensaje de confirmación.

Datos recopilados

Métricas

gcp.pubsub.snapshot.backlog_bytes
(gauge)
Total byte size of the messages retained in a snapshot.
Shown as byte
gcp.pubsub.snapshot.backlog_bytes_by_region
(gauge)
Total byte size of the messages retained in a snapshot, broken down by Cloud region.
Shown as byte
gcp.pubsub.snapshot.config_updates_count
(count)
Cumulative count of configuration changes, grouped by operation type and result.
Shown as update
gcp.pubsub.snapshot.num_messages
(gauge)
Number of messages retained in a snapshot.
Shown as message
gcp.pubsub.snapshot.num_messages_by_region
(gauge)
Number of messages retained in a snapshot, broken down by Cloud region.
Shown as message
gcp.pubsub.snapshot.oldest_message_age
(gauge)
Age (in seconds) of the oldest message retained in a snapshot.
Shown as second
gcp.pubsub.snapshot.oldest_message_age_by_region
(gauge)
Age (in seconds) of the oldest message retained in a snapshot, broken down by Cloud region.
Shown as second
gcp.pubsub.subscription.ack_latencies.avg
(gauge)
Average of ack latencies.
Shown as millisecond
gcp.pubsub.subscription.ack_latencies.samplecount
(count)
Sample Count for ack latencies.
Shown as millisecond
gcp.pubsub.subscription.ack_latencies.sumsqdev
(gauge)
Sum of Squared Deviation for ack latencies.
Shown as millisecond
gcp.pubsub.subscription.ack_message_count
(count)
Cumulative count of messages acknowledged by Acknowledge requests, grouped by delivery type.
Shown as byte
gcp.pubsub.subscription.backlog_bytes
(gauge)
Total byte size of the unacknowledged messages (also known as backlog messages) in a subscription.
Shown as byte
gcp.pubsub.subscription.billable_bytes_by_region
(gauge)
Total byte size of the billable messages in a subscription, broken down by Cloud region and type. See the Pub/Sub pricing page for more information about pricing for retaining unacknowledged messages in subscriptions for more than 24 hours.
Shown as byte
gcp.pubsub.subscription.byte_cost
(count)
Cost of operations per subscription measured.
Shown as byte
gcp.pubsub.subscription.config_updates_count
(count)
Number of configuration changes for subscriptions.
Shown as occurrence
gcp.pubsub.subscription.dead_letter_message_count
(count)
Cumulative count of messages published to dead letter topic.
Shown as message
gcp.pubsub.subscription.delivery_latency_health_score
(gauge)
A score that measures the health of a subscription.
Shown as occurrence
gcp.pubsub.subscription.exactly_once_warning_count
(count)
Count of instances (acks and modacks that may have failed) that may lead to message redeliveries.
Shown as occurrence
gcp.pubsub.subscription.expired_ack_deadlines_count
(count)
Cumulative count of messages whose ack deadline expired while the messages was outstanding to a subscriber client.
Shown as message
gcp.pubsub.subscription.export_push_subscription_details
(gauge)
Type and state of the export push subscription.
gcp.pubsub.subscription.mod_ack_deadline_message_count
(count)
Cumulative count of messages whose deadline was updated by ModifyAckDeadline requests.
Shown as message
gcp.pubsub.subscription.mod_ack_deadline_request_count
(count)
Number of ModifyAckDeadline requests.
Shown as request
gcp.pubsub.subscription.mod_ack_latencies.avg
(count)
The average distribution of ModAck latencies. For each message delivery attempt, the ModAck latency is the duration between when the CPS service delivers the message to a subscriber client and when the CPS service receives a ModAck request for that message (note that this means this metric may be recorded multiple times per delivery attempt). If this metric is growing, it indicates that the subscriber is repeatedly extending the acknowledgement deadline for a message it is holding.
Shown as millisecond
gcp.pubsub.subscription.mod_ack_latencies.samplecount
(count)
The sample count for distribution of ModAck latencies. For each message delivery attempt, the ModAck latency is the duration between when the CPS service delivers the message to a subscriber client and when the CPS service receives a ModAck request for that message (note that this means this metric may be recorded multiple times per delivery attempt). If this metric is growing, it indicates that the subscriber is repeatedly extending the acknowledgement deadline for a message it is holding.
Shown as millisecond
gcp.pubsub.subscription.mod_ack_latencies.sumsqdev
(count)
The sum of squared deviation for distribution of ModAck latencies. For each message delivery attempt, the ModAck latency is the duration between when the CPS service delivers the message to a subscriber client and when the CPS service receives a ModAck request for that message (note that this means this metric may be recorded multiple times per delivery attempt). If this metric is growing, it indicates that the subscriber is repeatedly extending the acknowledgement deadline for a message it is holding.
Shown as millisecond
gcp.pubsub.subscription.nack_requests
(gauge)
Cumulative count of messages that were nacked. One message nacked multiple times will be counted multiple times.
gcp.pubsub.subscription.num_outstanding_messages
(gauge)
Messages delivered but not yet acknowledged.
Shown as message
gcp.pubsub.subscription.num_retained_acked_messages
(gauge)
Number of acknowledged messages retained in a subscription.
Shown as message
gcp.pubsub.subscription.num_retained_acked_messages_by_region
(gauge)
Number of acknowledged messages retained in a subscription, broken down by Cloud region.
Shown as message
gcp.pubsub.subscription.num_unacked_messages_by_region
(gauge)
Number of unacknowledged messages in a subscription, broken down by Cloud region.
Shown as message
gcp.pubsub.subscription.num_undelivered_messages
(gauge)
Number of unacknowledged messages (also known as backlog messages) in a subscription.
gcp.pubsub.subscription.oldest_retained_acked_message_age
(gauge)
Age (in seconds) of the oldest acknowledged message retained in a subscription.
Shown as second
gcp.pubsub.subscription.oldest_retained_acked_message_age_by_region
(gauge)
Age (in seconds) of the oldest acknowledged message retained in a subscription, broken down by Cloud region.
Shown as second
gcp.pubsub.subscription.oldest_unacked_message_age
(gauge)
Age (in seconds) of the oldest unacknowledged message (also known as a backlog message) in a subscription.
Shown as second
gcp.pubsub.subscription.oldest_unacked_message_age_by_region
(gauge)
Age (in seconds) of the oldest unacknowledged message in a subscription, broken down by Cloud region.
Shown as second
gcp.pubsub.subscription.open_streaming_pulls
(gauge)
Number of open StreamingPull streams per subscription, grouped by status.
gcp.pubsub.subscription.pull_ack_request_count
(count)
Delta count of message pull acknowledgement requests.
Shown as request
gcp.pubsub.subscription.pull_request_count
(count)
Number of message pull requests.
Shown as request
gcp.pubsub.subscription.push_request_count
(count)
Number of message push attempts.
Shown as request
gcp.pubsub.subscription.push_request_latencies.avg
(gauge)
Average of push request latencies.
Shown as microsecond
gcp.pubsub.subscription.push_request_latencies.samplecount
(count)
Sample Count for push request latencies.
Shown as microsecond
gcp.pubsub.subscription.push_request_latencies.sumsqdev
(gauge)
Sum of Squared Deviation for push request latencies.
Shown as microsecond
gcp.pubsub.subscription.retained_acked_bytes
(gauge)
Total byte size of the acknowledged messages retained in a subscription.
Shown as byte
gcp.pubsub.subscription.retained_acked_bytes_by_region
(gauge)
Total byte size of the acknowledged messages retained in a subscription, broken down by Cloud region.
Shown as byte
gcp.pubsub.subscription.seek_request_count
(count)
Cumulative count of seek attempts, grouped by result.
Shown as request
gcp.pubsub.subscription.sent_message_count
(count)
Cumulative count of messages sent by Cloud Pub/Sub to subscriber clients.
Shown as message
gcp.pubsub.subscription.streaming_pull_ack_request_count
(count)
Cumulative count of streaming pull requests with non-empty acknowledge IDs, grouped by result.
Shown as request
gcp.pubsub.subscription.streaming_pull_mod_ack_deadline_request_count
(count)
Cumulative count of streaming pull requests with non-empty ModifyAckDeadline fields, grouped by result.
Shown as request
gcp.pubsub.subscription.streaming_pull_response_count
(count)
Cumulative count of streaming pull responses, grouped by result.
Shown as response
gcp.pubsub.subscription.unacked_bytes_by_region
(gauge)
Total byte size of the unacknowledged messages in a subscription, broken down by Cloud region.
Shown as byte
gcp.pubsub.topic.byte_cost
(count)
Byte cost of operations per topic.
Shown as byte
gcp.pubsub.topic.config_updates_count
(count)
Number of configuration changes for topics.
Shown as occurrence
gcp.pubsub.topic.ingestion_byte_count
(count)
Byte count of ingested messages per ingestion source type and import partition (for example, AWS Kinesis shard ID).
Shown as byte
gcp.pubsub.topic.ingestion_data_source_state
(gauge)
State of the ingestion data source per ingestion source type.
gcp.pubsub.topic.ingestion_failure_count
(count)
Number of errors encountered when ingesting data per ingestion source type, source name, import partition, and reason for error.
gcp.pubsub.topic.ingestion_message_count
(count)
Number of ingested messages per ingestion source type and import partition (for example, AWS Kinesis shard ID).
gcp.pubsub.topic.message_sizes.avg
(gauge)
Average of publish message sizes.
Shown as byte
gcp.pubsub.topic.message_sizes.samplecount
(count)
Sample Count for publish message sizes.
Shown as byte
gcp.pubsub.topic.message_sizes.sumsqdev
(gauge)
Sum of Squared Deviation for publish message sizes.
Shown as byte
gcp.pubsub.topic.num_retained_acked_messages_by_region
(gauge)
Number of acknowledged messages retained in a topic, broken down by Cloud region.
Shown as message
gcp.pubsub.topic.num_retained_messages
(gauge)
Number of messages retained in a topic.
Shown as message
gcp.pubsub.topic.num_unacked_messages_by_region
(gauge)
Number of unacknowledged messages in a topic, broken down by Cloud region.
Shown as message
gcp.pubsub.topic.oldest_retained_acked_message_age_by_region
(gauge)
Age (in seconds) of the oldest acknowledged message retained in a topic, broken down by Cloud region.
Shown as second
gcp.pubsub.topic.oldest_retained_message_age
(gauge)
Age (in seconds) of the oldest message retained in a topic.
Shown as second
gcp.pubsub.topic.oldest_unacked_message_age_by_region
(gauge)
Age (in seconds) of the oldest unacknowledged message in a topic, broken down by Cloud region.
Shown as second
gcp.pubsub.topic.retained_acked_bytes_by_region
(gauge)
Total byte size of the acknowledged messages retained in a topic, broken down by Cloud region.
Shown as byte
gcp.pubsub.topic.retained_bytes
(gauge)
Total byte size of the messages retained in a topic.
Shown as byte
gcp.pubsub.topic.schema_validation_latencies.avg
(count)
The average distribution of schema validation latencies in milliseconds.
Shown as millisecond
gcp.pubsub.topic.schema_validation_latencies.samplecount
(count)
The sample count for distribution of schema validation latencies in milliseconds.
Shown as millisecond
gcp.pubsub.topic.schema_validation_latencies.sumsqdev
(count)
The sum of squared deviation for distribution of schema validation latencies in milliseconds.
Shown as millisecond
gcp.pubsub.topic.send_request_count
(count)
Number of message send requests.
Shown as request
gcp.pubsub.topic.send_request_latencies.avg
(gauge)
Average of topic send request latencies.
Shown as microsecond
gcp.pubsub.topic.send_request_latencies.samplecount
(count)
Sample Count for topic send request latencies.
Shown as microsecond
gcp.pubsub.topic.send_request_latencies.sumsqdev
(gauge)
Sum of Squared Deviation for topic send request latencies.
Shown as microsecond
gcp.pubsub.topic.unacked_bytes_by_region
(gauge)
Total byte size of the unacknowledged messages in a topic, broken down by Cloud region.
Shown as byte

Eventos

La integración Google Cloud Pub/Sub no incluye eventos.

Checks de servicio

La integración Google Cloud Pub/Sub no incluye checks de servicio.

Resolución de problemas

¿Necesitas ayuda? Ponte en contacto con el servicio de asistencia de Datadog.