Integration troubleshooting

Start with the symptom below, then open Logs and debugging under Troubleshooting in the sidebar and copy build or deploy output before you contact support. The Troubleshooting overview under Troubleshooting in the sidebar links every major area.

For a shorter symptom list that also points here, see Integration issues under Troubleshooting in the sidebar.

Git: cannot see repositoriesCopy Anchor Link

1. Open Integrations and confirm the provider shows Connected.
2. GitHub: open the GitHub Settings for the installed app and verify the repo appears in Repository access. Add it if you use Only select repositories.
3. GitLab (connected OAuth): reconnect from Integrations and confirm the OAuth app includes api, read_user, and read_repository. For self-hosted, verify Test connection succeeds and the redirect URI matches your dashboard URL. See GitLab integration under Integrations in the sidebar.
4. Token providers (GitLab manual PAT, Bitbucket, Gitea, Azure DevOps): generate a new token with read scopes and reconnect. See Git source integrations.
5. Confirm you are in the right Organisation in dFlow. Integration visibility follows Tenant access model under Security and Team Management in the sidebar.

Git: clone or build fails with permission deniedCopy Anchor Link

1. Re-run connect or re-authorize the GitHub App if the org recently enabled SSO or third-party restrictions.
2. Confirm the branch exists and is spelled exactly as in the remote.
3. For private submodules or monorepos, ensure the same credentials can reach every URL the build touches (provider dependent).

If you suspect a dFlow role blocked an action another user performed, compare Roles and permissions under Security and Team Management in the sidebar.

Git: push does not trigger a new deployCopy Anchor Link

1. Check whether the Service uses manual deploy; push alone would not roll out.
2. GitHub App: confirm webhooks on the installation are healthy in GitHub Settings (delivery logs).
3. Confirm automatic deploy is supported for your provider and configuration (App services under Services in the sidebar).

Docker registry: pull denied or manifest unknownCopy Anchor Link

1. Open Integrations and edit the registry: re-enter the token, save, Redeploy the Docker Service.
2. Compare image:tag to the registry UI; typo or wrong digest causes manifest unknown.
3. For rate limits, authenticate with a user or token that has pull rights, then retry.

Guide: Docker registry integration.

Backup storage: test fails, unverified, or external backup errorsCopy Anchor Link

1. Open Integrations → Backup Storage and edit the provider: confirm bucket, region, and endpoint (blank for AWS S3 only).
2. Click Test connection with current keys; rotate keys in the cloud provider if the test fails.
3. Use Verify (shield icon) on the card if status is not verified before you select the provider on a database Backups tab.
4. If upload fails after a successful test, check bucket policy, object ownership, and that the database Service still has compute for SSH export. See Backups and restore under Databases in the sidebar.

Guide: Backup storage integration.

Azure: connection or provisioning errorsCopy Anchor Link

1. Open Integrations → Azure and confirm the account shows Connected. If not, click Connect account and sign in with Microsoft again.
2. Verify the signed-in identity still has access to the subscription and resource group you select in the worker form.
3. If lists load but VM creation fails, check Azure quotas (cores per region), VM size availability in the region, and whether the image is offered there.
4. For Advanced network security group mode, confirm your existing NSG allows inbound SSH (22), Tailscale (41641 UDP), monitoring (45876), HTTP (80), and HTTPS (443). Basic mode creates these rules for you.
5. If the worker is created in Azure but never shows Connected in dFlow, wait a few minutes, then open Compute troubleshooting under Worker Nodes and Compute in the sidebar.

Guide: Azure integration and Add an Azure server under Worker Nodes and Compute in the sidebar.

AWS: connection or provisioning errorsCopy Anchor Link

1. Verify Access key ID and Secret access key are current in IAM and in dFlow.
2. Open AWS integration and confirm the IAM policy includes the API actions the UI needs.
3. If EC2 starts but the server never registers, check Security groups, SSH, and Compute and worker node issues under Troubleshooting in the sidebar. Security groups under Security and Team Management in the sidebar explains network rules.

Mixed-up layers (Git vs SSH vs cloud)Copy Anchor Link

Still stuckCopy Anchor Link

1. Gather build log, deploy log, and a screenshot of the Integration status.
2. Read When to contact support under Troubleshooting in the sidebar.
3. Security-specific auth questions: Security troubleshooting under Security and Team Management in the sidebar.

RelatedCopy Anchor Link

• Integration permissions and prerequisites
• Deployment issues