Monitor Pub/Sub in Cloud Monitoring

Stay organized with collections Save and categorize content based on your preferences.

You can use the Google Cloud console or the Cloud Monitoring API to monitor Pub/Sub.

This document shows you how to to monitor your Pub/Sub usage in the Google Cloud console using Monitoring.

  • If you want to view metrics from other Google Cloud resources in addition to Pub/Sub metrics, use Monitoring.

  • Otherwise, you can use the monitoring dashboards provided within Pub/Sub. See Monitor topics and Monitor subscriptions instead of this document.

Before you begin

Before you use Monitoring, ensure that you've prepared the following:

  • A Cloud Billing account

  • A Pub/Sub project with billing enabled

One way to ensure that you've obtained both is to complete the Quickstart using the Cloud console.

View an existing dashboard

A dashboard lets you view and analyze data from different sources in the same context. Google Cloud provides both predefined and custom dashboards. For example, you can view a predefined Pub/Sub dashboard or create a custom dashboard that displays metric data, alerting policies, and log entries related to Pub/Sub.

To monitor your Pub/Sub project by using Cloud Monitoring, perform the following steps:

  1. In the Google Cloud console, go to the Monitoring page.

    Go to Monitoring

  2. Select the name of your project if it is not already selected at the top of the page.

  3. Click Dashboards from the navigation menu.

  4. In the Dashboards overview page, create a new dashboard or select the existing Pub/Sub dashboard.

    To search for the existing Pub/Sub dashboard, in the filter for All Dashboards, select the Name property and enter Pub/Sub.

For more information on how to create, edit, and manage a custom dashboard, see Manage custom dashboards.

View a single Pub/Sub metric

To view a single Pub/Sub metric by using the Google Cloud console, perform the following steps:

  1. In the Google Cloud console, go to the Monitoring page.

    Go to Monitoring

  2. In the navigation pane, select Metrics explorer.

  3. In the Configuration section, click Select a metric.

  4. In the filter, enter Pub/Sub.

  5. In Active resources, select Pub/Sub Subscription or Pub/Sub Topic.

  6. Drill down to a specific metric and click Apply.

    The page for a specific metric opens.

You can learn more about the monitoring dashboard by reading the Cloud Monitoring documentation.

View Pub/Sub metrics and resource types

Monitor quota usage

For a given project, you can use the IAM & Admin Quotas dashboard to view current quotas and usage.

You can view your historical quota usage by using the following metrics:

These metrics use the consumer_quota monitored resource type. For more quota-related metrics, see the Metrics list.

For example, the following Monitoring Query Language query creates a chart with the fraction of publisher quota being used in each region:

fetch consumer_quota
| filter resource.service == 'pubsub.googleapis.com'
| { metric serviceruntime.googleapis.com/quota/rate/net_usage
    | filter metric.quota_metric == 'pubsub.googleapis.com/regionalpublisher'
    | align delta_gauge(1m)
    | group_by [metric.quota_metric, resource.location],
        sum(value.net_usage)
  ; metric serviceruntime.googleapis.com/quota/limit
    | filter metric.quota_metric == 'pubsub.googleapis.com/regionalpublisher'
    | group_by [metric.quota_metric, resource.location],
        sliding(1m), max(val()) }
| ratio

If you anticipate your usage exceeding the default quota limits, create alerting policies for all the relevant quotas. These alerts fire when your usage reaches some fraction of the limit. For example, the following Monitoring Query Language query triggers an alerting policy when any Pub/Sub quota exceeds 80% usage:

fetch consumer_quota
| filter resource.service == 'pubsub.googleapis.com'
| { metric serviceruntime.googleapis.com/quota/rate/net_usage
    | align delta_gauge(1m)
    | group_by [metric.quota_metric, resource.location],
        sum(value.net_usage)
  ; metric serviceruntime.googleapis.com/quota/limit
    | group_by [metric.quota_metric, resource.location],
        sliding(1m), max(val()) }
| ratio
| every 1m
| condition gt(val(), 0.8 '1')

For more customized monitoring and alerting on quota metrics, see Using quota metrics.

See Quotas and limits for more information about quotas.

Maintain a healthy subscription

To maintain a healthy subscription, you can monitor several subscription properties using Pub/Sub-provided metrics. For example, you can monitor the volume of unacknowledged messages, the expiration of message acknowledgment deadlines, and so on. You can also check whether your subscription is healthy enough to achieve a low message delivery latency.

Refer to the next sections to get more details about the specific metrics.

Monitor message backlog

To ensure that your subscribers are keeping up with the flow of messages, create a dashboard. The dashboard can show the following backlog metrics, aggregated by resource, for all your subscriptions:

Create alerting policies that trigger when these values are outside of the acceptable range in the context of your system. For instance, the absolute number of unacknowledged messages is not necessarily meaningful. A backlog of a million messages might be acceptable for a million message-per-second subscription, but unacceptable for a one message-per-second subscription.

Common backlog issues

Symptoms Problem Solutions
Both the oldest_unacked_message_age and num_undelivered_messages are growing in tandem. Subscribers not keeping up with message volume
  • Add more subscriber threads or processes.
  • Add more subscriber machines or containers.
  • Look for signs of bugs in your code that prevent it from successfully acknowledging messages or processing them in a timely fashion. See Monitoring ack deadline expiration.
If there's a steady, small backlog size combined with a steadily growing oldest_unacked_message_age, there may be a few messages that cannot be processed. Stuck messages
  • Examine your application logs to understand whether some messages are causing your code to crash. It's unlikely—but possible —that the offending messages are stuck in Pub/Sub rather than in your client. Raise a support case after you're confident your code successfully processes each message.
  • If some messages are causing your code to crash, consider forwarding those messages to a dead-letter topic.
The oldest_unacked_message_age exceeds the subscription message retention duration. Permanent data loss
  • Set up an alert that fires before the lapsing of the message retention duration.

Monitor delivery latency health

In Pub/Sub, the delivery latency is the amount of time elapsed after a message is published and then delivered to a subscriber. If your message backlog is increasing, you can use the Delivery latency health score (subscription/delivery_latency_health_score) to check which factors are contributing to an increased latency.

This metric measures the health of a single subscription over a rolling 10-minute window. The metric provides insight into the following criteria, which are necessary for a subscription to achieve consistent low latency:

  • Negligible seek requests.

  • Negligible negatively acknowledged messages (nacked) messages.

  • Negligible expired message acknowledgment deadlines.

  • Consistent acknowledgment latency less than 30 seconds.

  • Consistent low utilization, meaning that the subscription consistently has adequate capacity to process new messages.

The Delivery latency health score metric reports a score of either 0 or 1 for each of the specified criteria. A score of 1 denotes a healthy state and a score of 0 denotes an unhealthy state.

  • Seek requests: If the subscription had any seek requests in the last 10 minutes, the score is set to 0. Seeking a subscription might cause old messages to be replayed long after they were first published, giving them an increased delivery latency.

  • Negatively acknowledged (nacked) messages: If the subscription had any negative acknowledgment (nack) requests in the last 10 minutes, the score is set to 0. A negative acknowledgment causes a message to be redelivered with an increased delivery latency.

  • Expired acknowledgment deadlines: If the subscription had any expired acknowledgment deadlines in the last 10 minutes, the score is set to 0. Messages whose acknowledgment deadline expired are redelivered with an increased delivery latency.

  • Acknowledgment latencies: If the 99.9th percentile of all acknowledgment latencies over the past 10 minutes was ever greater than 30 seconds, the score is set to 0. A high acknowledgment latency is a sign that a subscriber client is taking an abnormally long time to process a message. This score could imply a bug or some resource constraints on the subscriber client side.

  • Low utilization: Utilization is calculated differently for each subscription type.

    • StreamingPull: If you do not have enough streams open, the score is set to 0. Open more streams to ensure you have adequate capacity for new messages.

    • Push: If you have too many messages outstanding to your push endpoint, the score is set to 0. Add more capacity to your push endpoint so you have capacity for new messages.

    • Pull: If you do not have enough outstanding pull requests, the score is set to 0. Open more concurrent pull requests to ensure you're ready to receive new messages.

To view the metric, in Metrics explorer, select the Delivery latency health score metric for the Pub/Sub subscription resource type. Add a filter to select just one subscription at a time. Select the Stacked area chart and point to a specific time to check the criteria scores for the subscription for that point in time.

The following is a screenshot of the metric plotted for a one-hour period using a stacked area chart. The combined health score goes up to 5 at 4:15 AM, with a score of 1 for each criterion. Later, the combined score decreases to 4 at 4:20 AM, when the utilization score drops down to 0.

Screenshot of the delivery latency metric

Monitoring Query Language provides an expressive, text-based interface to Cloud Monitoring time-series data. The following MQL query creates a chart to measure the delivery latency health score for a subscription.

fetch pubsub_subscription
| metric 'pubsub.googleapis.com/subscription/delivery_latency_health_score'
| filter (resource.subscription_id == '$SUBSCRIPTION')
| group_by 1m,
   [value_delivery_latency_health_score_sum:
     sum(if(value.delivery_latency_health_score, 1, 0))]
| every 1m

Monitor acknowledgment deadline expiration

In order to reduce message delivery latency, Pub/Sub allows subscriber clients a limited amount of time to acknowledge (ack) a given message. This time period is known as the ack deadline. If your subscribers take too long to acknowledge messages, the messages are redelivered, resulting in the subscribers seeing duplicate messages. This redelivery can happen for various reasons:

  • Your subscribers are under-provisioned (you need more threads or machines).

  • Each message takes longer to process than the message acknowledgment deadline. Cloud Client Libraries generally extend the deadline for individual messages up to a configurable maximum. However, a maximum extension deadline is also in effect for the libraries.

  • Some messages consistently crash the client.

You can measure the rate at which subscribers miss the ack deadline. The specific metric depends on the subscription type:

Excessive ack deadline expiration rates can result in costly inefficiencies in your system. You pay for every redelivery and for attempting to process each message repeatedly. Conversely, a small expiration rate (for example, 0.1–1%) might be healthy.

Monitor message throughput

Pull and StreamingPull subscribers might receive batches of messages in each pull response; push subscriptions receive a single message in each push request. You can monitor the batch message throughput being processed by your subscribers with these metrics:

You can monitor the individual or unbatched message throughput being processed by your subscribers with the metric subscription/sent_message_count filtered by the delivery_type label.

Monitor push subscriptions

For push subscriptions, monitor these metrics:

  • subscription/push_request_count

    Group the metric by response_code and subcription_id. Since Pub/Sub push subscriptions use response codes as implicit message acknowledgments, it's important to monitor push request response codes. Because push subscriptions exponentially back off when they encounter timeouts or errors, your backlog can grow quickly based on how your endpoint responds.

    Consider setting an alert for high error rates since these rates lead to slow delivery and a growing backlog. You can create a metric filtered by response class. However, push request counts are likely to be more useful as a tool for investigating the growing backlog size and age.

  • subscription/num_outstanding_messages

    Pub/Sub generally limits the number of outstanding messages. Aim for fewer than 1,000 outstanding messages in most situations. After the throughput achieves a rate on the order of 10,000 messages per second, the service adjusts the limit for the number of outstanding messages. This limitation is done in increments of 1,000. No specific guarantees are made beyond the maximum value, so 1,000 outstanding messages is a good guide.

  • subscription/push_request_latencies

    This metric helps you understand the response latency distribution of the push endpoint. Because of the limit on the number of outstanding messages, endpoint latency affects subscription throughput. If it takes 100 milliseconds to process each message, your throughput limit is likely to be 10 messages per second.

To access higher outstanding message limits, push subscribers must acknowledge more than 99% of the messages that they receive.

You can calculate the fraction of messages that subscribers acknowledge using the Monitoring Query Language. The following MQL query creates a chart with the fraction of messages that subscribers acknowledge on a subscription:

fetch pubsub_subscription
| metric 'pubsub.googleapis.com/subscription/push_request_count'
| filter
    (resource.subscription_id == '$SUBSCRIPTION')
    | filter_ratio_by [], metric.response_class == 'ack'
    | every 1m

Monitor subscriptions with filters

Pub/Sub automatically acknowledges the messages that don't match a filter. You can monitor the number, size, and cost of these messages.

To monitor the number of messages that don't match a filter, use the subscription/ack_message_count metric with the delivery_type label and the filter value.

To monitor the size and cost of messages that don't match a filter, use the subscription/byte_cost metric with the operation_type label and the filter_drop value. For more information about the fees for these messages, see the Pub/Sub pricing page.

Monitor forwarded undeliverable messages

To monitor undeliverable messages that Pub/Sub forwards to a dead-letter topic, use the subscription/dead_letter_message_count metric. This metric shows the number of undeliverable messages that Pub/Sub forwards from a subscription.

To verify that Pub/Sub is forwarding undeliverable messages, you can compare the subscription/dead_letter_message_count metric with the topic/send_request_count metric. Do the comparison for the dead-letter topic to which Pub/Sub forwards these messages.

You can also attach a subscription to the dead-letter topic and then monitor the forwarded undeliverable messages on this subscription using the following metrics:

Maintain a healthy publisher

The primary goal of a publisher is to persist message data quickly. Monitor this performance using topic/send_request_count , grouped by response_code. This metric gives you an indication of whether Pub/Sub is healthy and accepting requests.

A background rate of retryable errors (lower than 1%) is not a cause for concern, since most Cloud Client Libraries retry message failures. Investigate error rates that are greater than 1%. Because non-retryable codes are handled by your application (rather than by the client library), you should examine response codes. If your publisher application does not have a good way of signaling an unhealthy state, consider setting an alert on the topic/send_request_count metric.

It's equally important to track failed publish requests in your publish client. While client libraries generally retry failed requests, they do not guarantee publication. Refer to Publishing messages for ways to detect permanent publish failures when using Cloud Client Libraries. At a minimum, your publisher application must log permanent publish errors. If you log those errors to Cloud Logging, you can set up a logs-based metric with an alerting policy.

Monitor message throughput

Publishers may send messages in batches. You can monitor the message throughput being sent by your publishers with these metrics:

  • topic/send_request_count: the volume of batch messages being sent by publishers.

  • A count of topic/message_sizes: the volume of individual (unbatched) messages being sent by publishers.

    You can calculate a count of messages being sent by applying a count aggregator to this metric, or by using the Monitoring Query Language. The following MQL query creates a chart with the rate of individual messages sent on a topic:

    fetch pubsub_topic
    | metric 'pubsub.googleapis.com/topic/message_sizes'
    | filter
        (resource.topic_id == '$TOPIC')
    | align delta(1m)
    | every 1m
    | group_by [], [row_count: row_count()]
    

What's next