Ralf Santa Cruz
12/04/2025, 12:25 AMkapa.ai
12/04/2025, 12:25 AMDATA_PLANE_CLIENT_ID - DATA_PLANE_CLIENT_SECRET inside the airbyte-config-secrets secret in the data-plane cluster. [[Multiple region](https://docs.airbyte.com/platform/1.8/enterprise-setup/multi-region#step-4)] Example: yaml apiVersion: v1 kind: Secret metadata: name: airbyte-config-secrets type: Opaque data: # Insert the data plane credentials received in step 2 DATA_PLANE_CLIENT_ID: your-data-plane-client-id DATA_PLANE_CLIENT_SECRET: your-data-plane-client-secret On the control‑plane side, there is also a dataplane / dataplane_client_credentials table in the database that stores these credentials; deleting those rows forces regeneration in some versions, as seen in the 1.6.0 issue. [[GitHub 58587](https://github.com/airbytehq/airbyte/issues/58587)] ### 2. How credentials may be regenerated (what is known) From the 1.6.0 workaround: 1. Delete rows from dataplane_client_credentials and dataplane in the control-plane DB, then 2. Re-run abctl install, which caused the dataplane and its credentials to be recreated and written into the appropriate secrets. [[GitHub 58587](https://github.com/airbytehq/airbyte/issues/58587)] That issue is for 1.6.0, not 1.8.x. The knowledge sources do not explicitly say this is the supported or recommended way for 1.8.x, nor do they document a UI/API “rotate data plane credentials” action. ### 3. What is not documented - There is no 1.8.x doc that describes a first‑class, supported workflow to “re‑issue / rotate” a data plane client ID and secret. - There is no doc stating that deleting the DB rows and re‑running abctl is safe or supported behavior for 1.8.x. - There is no doc showing a UI or public API endpoint for regenerating data plane credentials. Because of this, I can’t provide a definitive, version‑guaranteed procedure for 1.8.x without going beyond the docs. ### 4. Practical next step Given the lack of explicit 1.8.x guidance: - Treat data plane credential rotation as an Enterprise / multi‑region feature question rather than a general OSS one. - I strongly recommend opening a support ticket (if you’re on Self‑Managed Enterprise) or posting on the Airbyte community forum / GitHub with your exact 1.8.x patch version, and asking for the official rotation procedure for data plane credentials. Reference that: - Your data plane uses DATA_PLANE_CLIENT_ID and DATA_PLANE_CLIENT_SECRET in airbyte-config-secrets. [[Multiple region](https://docs.airbyte.com/platform/1.8/enterprise-setup/multi-region#step-4)] - There is historical guidance for 1.6.0 involving dataplane_client_credentials and dataplane tables, but nothing documented for 1.8.x. [[GitHub 58587](https://github.com/airbytehq/airbyte/issues/58587)] Right now, the knowledge sources don’t contain enough information to safely describe a concrete, supported rotation procedure for data plane credentials on Airbyte 1.8.x.Ralf Santa Cruz
12/04/2025, 12:39 AM│ 2025-12-04 00:35:02,169 [scheduled-executor-thread-1] WARN i.a.a.c.c.ClientConfigurationSupport(generateDefaultRetryPolicy$lambda$5):79 - Retry attempt 4 of 5. Last response: null │
│ 2025-12-04 00:35:03,186 [scheduled-executor-thread-1] ERROR i.a.a.c.a.AccessTokenInterceptor(intercept):120 - Failed to obtain or add access token │
│ java.net.ConnectException: Failed to connect to data-aws-rnd-airbyte-rnd-airbyte-server-svc/172.20.136.252:8001
my workload launcher is unable to hit the serverkapa.ai
12/04/2025, 12:39 AMkapa.ai
12/04/2025, 12:39 AMairbyte-server service: text Failed to connect to data-aws-rnd-airbyte-rnd-airbyte-server-svc/172.20.136.252:8001 This is the same class of error others hit when components cannot reach airbyte-...-server-svc or the workload API service. [[Issue 68134](https://github.com/airbytehq/airbyte/issues/68134); [Issue 68608](https://github.com/airbytehq/airbyte/issues/68608)] From the available information, the recommended checks are: 1. Verify the server service and pods are running bash kubectl get pods -n <your-namespace> kubectl get svc -n <your-namespace> | grep airbyte-server In similar cases, the problem was simply that the server couldn’t start or wasn’t reachable, causing all dependent components (workload launcher, connectors) to fail with ConnectException. [[Issue 68134](https://github.com/airbytehq/airbyte/issues/68134); [Issue 68608](https://github.com/airbytehq/airbyte/issues/68608)] 2. Check server pod logs Look at the logs for airbyte-airbyte-server (or equivalent) to see why it might be failing or crash‑looping: bash kubectl logs -n <your-namespace> <airbyte-server-pod> In one Helm deployment issue, server initialization errors prevented it from accepting connections, which then made the workload launcher fail in exactly this way. [[Issue 68134](https://github.com/airbytehq/airbyte/issues/68134)] 3. Confirm in‑cluster DNS and service name Ensure the workload launcher is trying to contact the correct service name and namespace (e.g. *-airbyte-server-svc in the same namespace). The errors in the issues show failures connecting to the *-server-svc ClusterIP; if the service name or namespace is wrong, you’ll see the same ConnectException. [[Issue 68134](https://github.com/airbytehq/airbyte/issues/68134); [Issue 68608](https://github.com/airbytehq/airbyte/issues/68608)] 4. If using abctl / separate workload API, ensure both sides are up For abctl‑based deployments, similar Failed to connect errors were resolved by restarting the worker and workload API deployments once configuration was fixed: bash kubectl -n airbyte-abctl rollout restart deployment airbyte-abctl-worker kubectl -n airbyte-abctl rollout restart deployment airbyte-abctl-workload-api-server kubectl -n airbyte-abctl rollout status deployment airbyte-abctl-worker kubectl -n airbyte-abctl rollout status deployment airbyte-abctl-workload-api-server [[Issue 66692](https://github.com/airbytehq/airbyte/issues/66692)] The knowledge sources do not provide a more specific fix for this exact hostname/namespace, so the next step is to inspect the airbyte-server pod’s status and logs; usually the root cause is there (crash, misconfig, missing secrets, etc.), and until that server comes up healthy, the workload launcher will keep failing with this ConnectException.