Configure Telemetry Data Transfer

The data plane (DP) periodically reports operational metrics to the control plane (CP) to enable centralized monitoring, alerting, and dashboard visualizations. This telemetry includes API traffic volumes, request latency, HTTP status codes, bandwidth, and connection counts in prometheus format.

By default, the DP reports telemetry data every 15 seconds with Gzip compression enabled. The control plane owns this configuration and sends it to every connected gateway group.

Configure Telemetry

Configure telemetry in the Dashboard configuration. The Dashboard stores these settings in the control-plane database and includes them in the configuration payload sent to connected data planes.

For Helm deployments, add the settings under dashboard_configuration in the values file for the api7/api7ee3 release:

cp-values.yaml

dashboard_configuration:
  telemetry:
    enable: true
    interval: 15
    max_metrics_size: 33554432
    metrics_batch_size: 4194304
    compression_level: -1

❶ Enables periodic metrics uploads to the control plane.

❷ Sets the minimum interval, in seconds, between upload attempts.

❸ Stops creating additional batches after the accumulated uncompressed metrics reach this threshold. Remaining metrics are truncated. The default is 32 MB.

❹ Sets the target uncompressed batch size for each upload request. Batches keep metric lines whole, so a batch can slightly exceed this target. This value cannot exceed max_metrics_size. The default is 4 MB.

❺ Sets the Gzip compression level from -1 through 9. The default -1 uses the compression library's default level.

For Docker Compose or package deployments, add the same telemetry block at the top level of the Dashboard configuration file:

Dashboard configuration

telemetry:
  enable: true
  interval: 15
  max_metrics_size: 33554432
  metrics_batch_size: 4194304
  compression_level: -1

Apply the control-plane configuration change and restart or roll out the Dashboard. The Dashboard updates the stored telemetry settings. When existing stored values change, it also advances every gateway-group configuration version so connected data planes receive the new settings through normal configuration synchronization. The data planes do not require a separate restart.

Configure every Dashboard replica with the same telemetry values. A replica that starts with different values can replace the stored global configuration.

Do not configure api7ee.telemetry only in a connected gateway's local config.yaml. The control-plane payload takes precedence over that local block, so the next configuration update can replace the local values.

Choose a Compression Level

Gzip compression reduces the network overhead of telemetry transfers. The compression level controls the trade-off between CPU usage and data size:

Level	Description	Characteristics
-1	Library default	Typically level 6 (balanced)
0	No compression	Minimal compression effort; largest payload
1	Fastest compression	Low compression CPU usage; low compression ratio
6	Balanced compression	Moderate compression CPU usage and payload size
9	Maximum compression	Highest compression effort; typically smallest payload

For most deployments, keep the default level -1. Select another level only after measuring gateway CPU usage and compressed payload sizes with representative configuration and traffic; the result depends on the volume and content of the collected metrics.

Disable Telemetry Uploads

To disable telemetry data transfer from every connected DP, set enable to false in the control-plane configuration:

cp-values.yaml

dashboard_configuration:
  telemetry:
    enable: false

For a file-based deployment, set the root Dashboard configuration instead:

Dashboard configuration

telemetry:
  enable: false

Apply the change and restart or roll out the Dashboard. Confirm that the connected gateway groups receive a new configuration version and that new traffic metrics stop appearing in the Dashboard. Disabling the upload does not disable gateway configuration synchronization or gateway health reporting. It also does not affect a separately enabled Prometheus endpoint. An external monitoring system can continue scraping the gateway only when that endpoint and its listener or Service exposure are configured and reachable.

Impacts of Disabling Telemetry

Disabling telemetry has significant operational impacts:

Impact	Details
Dashboard analytics	Traffic, latency, status-code, bandwidth, and connection views stop receiving new DP metrics.
Metric-based alerts	Alert policies that depend on the uploaded gateway metrics no longer receive new samples. Control-plane alerts based on other state can continue.
Operational visibility	Capacity planning and troubleshooting must use an alternative metrics path, such as direct Prometheus scraping.

Before disabling uploads, ensure that another monitoring path provides the visibility your operational and compliance requirements need.

note

These telemetry reports are sent only from the data plane to its configured control plane. This setting does not control other external integrations.

Next Steps

• Metrics: Configure gateway metrics collection and Prometheus exposure.
• Alerts: Configure alerts that use gateway and control-plane data.
• Control Plane Configuration: Review Dashboard and DP Manager configuration.