Title here
Summary here
Deployment
This page is a compact deployment reference for operators. It is designed for teams that must brief DSI/RSSI and run production environments.
Audience
- platform and DevOps teams deploying Fred,
- security/compliance stakeholders reviewing controls,
- run teams operating customer platforms.
Delivery And Promotion Model
Typical lifecycle:
- Features are merged on
develop. developis deployed to integration.- Validated changes are promoted to
main. mainis tagged (code/chart release tags).- Production is deployed from artifacts built from
main.
This operating model also applies to customer forks with their own CI/CD.
Fork Operating Model
For customer platforms:
- the customer forks Fred repositories,
- tracks upstream changes by controlled sync,
- keeps environment overlays (values, secrets integration, infra specifics) in customer repositories,
- promotes releases with the same governance (integration on
develop, production on taggedmain).
Deployment Building Blocks
A typical production setup includes:
- frontend,
- agentic-backend,
- knowledge-flow-backend,
- identity provider (Keycloak/OIDC),
- storage services (PostgreSQL, optional ClickHouse/object storage),
- Temporal server and one or several workers.
Temporal Position
Knowledge Flow processing is executed through Temporal workflows. Worker count is a scaling lever and should be part of capacity planning.
Platform Responsibilities
Before go-live, platform teams usually provide:
- Kubernetes namespaces, ingress, certificates, DNS.
- Data services and backup policy.
- Temporal deployment and scaling baseline.
- Keycloak realm/clients/groups/roles bootstrap.
- Secret management and rotation pipeline.
- Observability baseline (logs, metrics, traces, alerts).
Day-2 Operations
- Track release/tag promotions.
- Monitor queue health and worker saturation.
- Review security posture and credential rotation.
- Review storage growth and retention.
- Keep overlays aligned with upstream changes.