A missed credential can stop your observability pipeline cold. One bad secret rotation and your dashboard reads like static. That is where Dynatrace and Google Cloud Secret Manager make a surprisingly disciplined duo. Together they keep metrics flowing while your keys stay out of human hands.
Dynatrace tracks everything that moves. GCP Secret Manager stores everything you should never see. When you connect them, Dynatrace can ingest configuration values, tokens, or service credentials without exposing them in plain text or code. It is the cleanest form of trust boundary: data visibility where it belongs, secrets where they don’t.
The workflow is straightforward. Dynatrace agents or extensions reference secret versions stored in GCP Secret Manager using IAM permissions tied to a service account. That service account maps through Google Cloud IAM with roles fine-tuned for read-only access. Dynatrace pulls the secret at runtime, decrypts it in memory, and runs its checks or integrations. No local secret files, no shared credentials, no forgotten rotation scripts.
When setting up, think permissions before automation. Assign least-privilege roles. Handle key rotation with version IDs instead of overwriting existing secrets. Log access with Cloud Audit Logs to trace which process viewed what. A missed permission can break ingestion, but a sloppy one can leak credentials. Keep the principle simple: if the instance does not need to know, it should not even ask.
Quick setup answer: Dynatrace connects to GCP Secret Manager through a service account using the Secret Manager API. You grant read access to that account and reference the stored secret in Dynatrace configuration. Rotation is handled by updating secret versions, so credentials stay current without redeploying agents.