Security, credentials and governance for orchestration
An orchestrator needs permission to start jobs in every platform it coordinates, which makes it one of the most privileged tools in a data estate. The governance questions are therefore simple to ask: where are the credentials kept, what can each identity do, who can sign in, where is the configuration stored, and who has audited it.
Questions to ask any orchestration tool
| Question | Why it matters | Polysync's answer |
|---|---|---|
| Where are platform credentials stored? | A copy in the tool's database is a second place to leak them | In your own vault (Azure Key Vault, AWS Secrets Manager, Google Cloud Secret Manager or HashiCorp Vault), read at dispatch time, or in Polysync's managed Azure Key Vault if you prefer |
| How are secrets referenced? | Values typed into job parameters can end up in logs and exports | By name (vault://vault/secret), never stored on the task |
| How do users sign in? | Separate passwords escape your MFA and leaver process | Microsoft Entra ID work accounts, so your MFA and conditional access apply |
| How is customer data separated? | Shared tables depend on every query filtering correctly | Each customer has its own database schema |
| Where is data held? | Residency rules for configuration and run history | Azure Australia East, with backups in a second Australian region; AI requests are processed in Australia East |
| Has anyone audited it? | A vendor questionnaire is not evidence | Microsoft 365 Certification, renewed annually |
Least privilege per platform
Give the orchestrator's identity the narrowest role the platform offers for listing, starting, polling and cancelling jobs, or a custom role with only those actions where the built-in roles are broader. Each Polysync platform page has a permissions checklist, for example Azure Data Factory (Data Factory Contributor, or a custom role with equivalent pipeline-run permissions) and Azure Databricks.
Frequently asked questions
Does Polysync see my pipeline data?
No. Pipelines, notebooks and models keep running in your own cloud accounts. Polysync stores the configuration and run records needed to orchestrate them, and small parameter values you choose to pass between tasks.
Can I use my own Azure Key Vault?
Yes. Link the vault once and reference secrets by name from any platform connection. Polysync reads the secret when a task is dispatched.
Is Polysync hosted in Australia?
Yes. It runs in Azure's Australia East region, stores configuration and run history in Australia, and keeps backups in a second Australian region.
Answers in this topic
- How can an orchestration tool use my secrets without storing my credentials?Reference secrets by name from your own vault, resolved at dispatch time, instead of pasting them into the tool.