Self-monitoring for the Dynatrace GCP integration

  • How-to guide
  • 4-min read
  • Published Jul 16, 2021

Self-monitoring allows quick diagnosis to determine if a self-monitoring function is properly processing and sending logs to Dynatrace

Enable self-monitoring

Follow the steps below according to your deployment scenario.

Enable self-monitoring for the GKE deployment

  1. Connect to the Kubernetes cluster where the GCP Monitor deployment is running.

  2. Edit the configmap.

    kubectl -n dynatrace edit configmaps dynatrace-gcp-monitor-config

  3. Change the value of the SELF_MONITORING_ENABLED parameter to true.

  4. Restart the GKE GCP Monitor.

    kubectl -n dynatrace rollout restart deployment dynatrace-gcp-monitor

Enable self-monitoring for the GCP Monitor deployment

  1. In the Google Cloud console, go to Cloud Functions.
  2. Select dynatrace-gcp-monitor.
  3. Select Edit.
  4. Under Runtime, build and connection settings, change the value of the SELF_MONITORING_ENABLED runtime environment variable to true.
  5. Select Next, and then select Deploy to apply the new settings.

Self-monitoring metrics

The Dynatrace GCP Monitor deployment reports self-monitoring metrics as Google Cloud metrics. See below the list of self-monitoring metrics for metric/log ingest.

Self-monitoring metrics for the GKE deployment

Metric ingestion

Metric
Description
MINT lines ingested
The number of data points (metrics with dimensions) ingested by Dynatrace Metrics API v2 in a given interval.
Dynatrace connectivity
The connectivity status (1 = OK) between the monitoring function and Dynatrace. Connectivity can be broken due to an incorrect Dynatrace URL, an incorrect API token, or network connectivity issues.
Dynatrace failed requests count
The number of requests rejected by Dynatrace Metrics API v2. The reason for failure can be that the data point value doesn't comply with the metric ingestion protocol syntax, or that the limit for metric ingestion is exceeded.
Dynatrace requests count
The number of requests sent to Dynatrace.

Log ingestion

Metric
Description
All requests
All requests sent to Dynatrace
Dynatrace connectivity failures
The number of failed requests to connect to Dynatrace
Too old records
The number of log records that were invalid because the timestamp was too old
Too long content size
The number of records with content exceeding the maximum content length
Parsing errors
The number of errors that occurred while parsing logs
Processing time
The total amount of time for processing logs
Sending time
The total amount of time for sending logs
Sent logs entries
The number of log entries sent to Dynatrace
Log ingest payload size
The size of the log payload sent to Dynatrace (in KB)

Self-monitoring metrics for the GCP Monitor deployment

Metric ingestion

Metric
Description
MINT lines ingested
The number of data points (metrics with dimensions) ingested by Dynatrace Metrics API v2 in a given interval.
Dynatrace connectivity
The connectivity status (1 = OK) between the monitoring function and Dynatrace. Connectivity can be broken due to an incorrect Dynatrace URL, an incorrect API token, or network connectivity issues.
Dynatrace failed requests count
The number of requests rejected by Dynatrace Metrics API v2. The reason for failure can be that the data point value doesn't comply with the metric ingestion protocol syntax, or that the limit for metric ingestion is exceeded.
Dynatrace requests count
The number of requests sent to Dynatrace.

View self-monitoring metrics

The self-monitoring dashboards present multiple metrics related to Dynatrace connectivity status, the amount of data processed, and execution times.

To view the dashboards with self-monitoring metrics

  1. In your GCP console, go to the GCP Monitoring service.
  2. Select Dashboards.
  3. Depending on the type of deployment selected, search for
    • The dynatrace-gcp-monitor log self monitoring dashboard (for logs)
    • The dynatrace-gcp-monitor self monitoring dashboard (for metrics)

Example dashboard:

Self monitor

Related tags
Infrastructure Observability