Administration runbook

Use this runbook after Gluesync is installed. It links each routine or high-risk operation to the procedure that owns it.

Daily checks

  1. Open Pipelines and look for On error or pipelines that remain On warm-up.

  2. Open the Notifications Hub and review unread errors and warnings.

  3. Check entity latency, throughput, and buffer pressure for important pipelines.

  4. Confirm that source and target connectors are connected.

Start with Pipeline states and error recovery when a pipeline is unhealthy. Use Control Plane monitoring for metric definitions.

Changes to pipeline configuration

  1. Take a configuration backup.

  2. Put the affected pipeline in maintenance mode if the change requires connector reconnection or schema editing.

  3. Make the change.

  4. Exit maintenance mode.

  5. Verify connector health and test representative entities.

Pausing an entity and entering maintenance mode are different operations. Pause for replication control; use maintenance mode for configuration work that disconnects the pipeline.

Updates

  1. Review the release notes for the target version.

  2. Confirm that the latest configuration backup can be restored.

  3. Schedule an operational window.

  4. Open Platform updates in the Control Plane.

  5. Update services and retry only the service that fails.

  6. Verify pipeline and connector health after the update.

See Update Gluesync for release channels, update states, and failure handling. Air-gapped deployments must use the offline update procedure.

Backup and recovery

Back up Core Hub configuration on a schedule and before every platform or pipeline change. Store the exported bundle outside the Gluesync host and protect it as a sensitive configuration artifact.

At least once per release cycle:

  1. Restore the latest backup into a non-production environment.

  2. Re-supply secrets that are intentionally omitted or masked.

  3. Confirm that pipelines, connectors, entities, and UDFs are present.

  4. Record the restore duration and any manual step.

See Backup and restore.

Users, tokens, and identity

Use Settings for local users, roles, personal API tokens, and OIDC configuration.

  • Review privileged accounts regularly.

  • Remove unused tokens and users.

  • Prefer named users or service-specific tokens over shared credentials.

  • Test OIDC changes in a recovery-safe window and retain a documented local-admin recovery path.

See Settings and account and User management and RBAC.

Logs and support

For an incident:

  1. Record the affected pipeline, entity, connector, and time range.

  2. Preserve the first error; later errors are often consequences.

  3. Use Settings > Support to collect a diagnostic archive.

  4. Include reproduction steps and recent configuration or update changes.

  5. Upload the archive through the support workflow or attach it to the ticket.

See Enterprise support and Troubleshooting.

Production controls

Control Procedure

Metrics and dashboards

Monitoring

Email, webhooks, and smart alerts

Alerting

Log storage and retention

Logging

TLS, node encryption, and database transport

Secure communications

Sizing and connection tuning

Operational best practices

Production-readiness review

Move to production