You can use the Google Cloud console to access the monitoring dashboardfor a single topic or all the topics available in your project.

• A Cloud Billing account

• A Pub/Sub project with billing enabled

• At least one topic in your Pub/Sub project

• Familiarity withCloud MonitoringandMetrics Explorer

View the monitoring dashboard for all topics

To view the monitoring dashboard for all the topicsin a project, perform the following steps:

1. In the Google Cloud console, go to the Pub/SubTopics page.

   Go to Topics

2. Click theMetrics tab.

   The monitoring dashboard is displayed for all topics in the project.

The monitoring dashboard has two sections:Overview andQuota.

View the monitoring dashboard for a single topic

To view the monitoring dashboard for a single topicin a project, perform the following steps:

1. In the Google Cloud console, go to the Pub/SubTopics page.

   Go to Topics

2. In theTopics page, click any topic.

   The topic details page is displayed.

3. Within the topic details page, click theMetrics tab.

   The monitoring dashboard is displayed for the single topic.

The monitoring dashboard has three sections:Overview,Subscriptions,andRetention. You can see theRetention section only if youenable message retention for the topic. For more information about messageretention, seeProperties of a topic.

Customize the monitoring dashboard

You can create a custom dashboard or add individual charts to anexisting custom dashboard.

To create a custom dashboard, perform the following steps:

1. Follow the steps to go to the monitoring page forall topics or forasingle topic.

2. Choose one of the following options:

   • To create a new dashboard containing all the charts, or to add all thecharts to an existing custom dashboard, clickSave as custom dashboard.

   • To add a single chart to a new or custom dashboard, in an individual chart,expand themore_vertMore Actions option and clickAdd to custom dashboard.

You can view all your dashboards in theMonitoring console.

Modify a chart

To modify a chart, perform the following steps:

1. Follow the steps to go to the monitoring page forall topics or forasingle topic.

2. In an individual chart, expand themore_vertMore Actions option and clickView in Metrics Explorer.

   The chart opens inMetrics Explorer.

3. You can usePrometheus Query Language (PromQL) queries to make changes to yourchart and test the output.

4. ClickSave chart to save the changes.

Create an alerting policy

Creating an alerting policy for a chart lets you set up an alert monitoringmetric in the chart, configure the circumstances under which you want to bealerted, and configure how you want to be notified.

For more information about how to create alerts, seeCreate metric-based alert policies.

Some charts might not offer aCreate alerting policy option.In this case, perform the following steps:

1. Save the chart to a custom dashboard.

2. In the custom dashboard, for the individual chart, expandthemore_vertMore Actions option.

3. ClickConvert to alert chart.

Key charts to monitor topics

Although there are a number of charts available to monitor a single topic,the following charts in theOverview section are a good starting point:

• ThePublish requests chart lets you check if there are sudden spikesin the topic for receiving publish requests.

• TheAverage number of messages per batch chart lets you checkif the batching size is adequate for your requirements. A smaller batch sizeleads to reduced publish latency.

The following charts in theQuota section for all topics help you monitor quotasfor all the topics in your project:

• TheAdministrator operations chart lets you monitor administrativeoperations regarding topics. Each administrator operation charges oneunit against this quota.

For more information, seePub/Sub quotas and limits.

Metrics to monitor import topics

You can use the following metrics to monitor import topics:

• Ingestion byte count:This metric displays thetotal volume of data (in bytes) ingested from an import source into animport topic. It provides a breakdown of byte counts for each importsource and partition (shard).

• Ingestion message count:This metric displaysthe total number of import messages for each import source and partition.

• Ingestion data source state:This metricdisplays the state of the external source from which the import topic isingesting data. The error codes and their workarounds are listed in thetroubleshooting section of the import topics.For more information, see the following:

  • Troubleshooting Amazon Kinesis Data Streams import topics.
  • Troubleshooting Cloud Storage import topics.
  • Troubleshooting Amazon MSK import topics.
  • Troubleshooting Azure Event Hubs import topics.
  • Troubleshooting Confluent Cloud import topics.

• Ingestion failure count:This metric displays the number of failures encountered during ingestion.For Cloud Storage import topics, this metric includes the bucketname and the failure reason. For Amazon MSK, Azure Event Hubs, and ConfluentCloud import topics, this metric includes the Kafka topic or event hub name,the partition, and the failure reason.

To check these metrics, follow the steps described inView the monitoring dashboard for a single topic.

What's next

• To monitor subscriptions,seeMonitor subscriptions in Pub/Sub.

• To learn more about using PromQL to buildmonitoring charts, seeUsing the Query Editor.