GCP connection management overview
- Latest Dynatrace
- Explanation
A connection can see multiple Google Cloud projects. Use the Projects tab to view which projects are visible and monitored.
If the same project is visible to multiple connections, you risk ingesting duplicate data. Dynatrace detects and flags these situations. To resolve the issue, either change the service account permissions or use the connection filters—see Duplicate data from overlapping connections for the step-by-step guide.
Connection health
The Overview tab shows ingestion status for topology, metrics, and logs.
The Health tab displays connection status. A new connection starts as Pending until Dynatrace successfully authenticates with your Google Cloud service account. A connection is Unhealthy if authentication or Google Cloud access fails.
| Status | Meaning |
|---|---|
Pending | Connection created in Dynatrace, but Google Cloud setup is not yet complete or authentication has not succeeded. |
Healthy | Dynatrace is successfully authenticating. A Healthy status confirms authentication succeeded but does not guarantee data is flowing—if topology, metrics, or logs are missing, check the Overview tab ingestion status or see Troubleshoot GCP connections. |
Unhealthy | Authentication or data collection is failing. Check the Health tab for details. |
Operations
Once a connection is Healthy, you can inspect and observe it in Smartscape on Grail with DQL—see what it monitors and track its ingestion.
Connection topology entity
Each connection is represented in Smartscape on Grail as a DA_MONITORING_CONFIGURATION entity. The entity records which Google Cloud projects the connection sees (in visibility scope) and monitors (actively collecting), together with the linked service account and configuration ID.
-
List your Google Cloud connection entities:
smartscapeNodes "DA_MONITORING_CONFIGURATION"| filter cloud.provider == "gcp"| fields id, cloud.acquisition.status, dt.extension.config.id, gcp.service_account.email -
Show which projects each connection sees and monitors:
smartscapeEdges "*"| filter source_type == "DA_MONITORING_CONFIGURATION" and startsWith(target_type, "GCP")| lookup [smartscapeNodes "GCP*" | fieldsAdd project = gcp.project.id | fields id, project],sourceField:target_id, lookupField:id, prefix:"p."| fields connection = source_id, relation = type, project = p.project
- monitors
The connection actively collects data from the project.
- sees
The project is visible to the connection but may be excluded by folder, project, label, or tag filters.
Connection metrics
Each connection also exposes self-monitoring metrics, dimensioned by the configuration ID (dt.extension.config.id):
| Metric | Description |
|---|---|
dt.sfm.da.gcp.metric.data_points.count | Cloud Monitoring metric data points ingested |
dt.sfm.da.gcp.log.records.count | Log records forwarded |
dt.sfm.da.gcp.smartscape.updates.count | Topology (Smartscape) updates applied |
dt.sfm.extension.config.status | Monitoring configuration status |
dt.sfm.extension.status | Extension runtime status |
Query ingested metric data points per connection over the last two hours:
timeseries points = sum(dt.sfm.da.gcp.metric.data_points.count),by:{ dt.extension.config.id, gcp.project.id }, from:-2h
Connection self-monitoring metrics link to the connection entity via dt.extension.config.id. Ingested resource metrics (cloud.gcp.*) instead link to the Smartscape entity they describe via dt.smartscape_source.id—see GCP topology.
Update a connection
After a connection is created, you manage and update its configuration—monitoring settings, scope, log forwarding, asset feed, token rotation, and more—using one of three methods: the Settings app, Terraform, or the gcloud CLI with dtctl. Not every task is supported by every method. Use the table to find the method for the task you need, then follow the linked steps.
If you deployed with Terraform, keep two update tracks in sync:
- Terraform bundle (GCP infrastructure)
- Extension (Dynatrace-side monitoring engine)
They are updated independently.
Troubleshooting
If a connection is Unhealthy, or topology, metrics, or logs are missing, see Troubleshoot GCP connections. It covers connection status (Pending/Unhealthy), symptom-based fixes, and a step-by-step data-flow checklist.