Glossary

Glossary

Single-line definitions for every Styrmin term that appears in the docs, the GraphQL schema, the SDK, the CLI, and the UI. Use this page when you encounter a word and want to know what it means in one sentence. For the mental model and how the pieces fit together, read Core concepts.

A

Action — A Python function declared by an Application Driver that runs inside a Prefect flow at a specific lifecycle hook and mode (e.g. setup / pre). See Lifecycle hooks and actions.

Active instance — The single Instance currently considered live for a Deployment. Reconciliation compares the newly generated Instance against the active Instance.

Agent — The container Styrmin runs inside a target Kubernetes cluster. Reconciles recorded intent into Kubernetes resources and reports observed state back to the server. One per cluster.

Application Driver — A versioned, self-contained deployment specification for one application. The unit of deployment in Styrmin. See What is an Application Driver?.

Application Driver Version — A specific revision of an Application Driver, with its own Helm chart pin, components, parameters, services, and actions. A Deployment points at one Application Driver Version.

Application Version — The user-facing version of the application itself (e.g. Infrahub 1.6.0). Distinct from the Application Driver Version that knows how to deploy it.

B

Backup Storage Location (BSL) — An S3-compatible object store registered as a destination for Velero backups. Assigned to an Environment; all deployments in that environment back up there.

C

CLI / styrminctl — The Cyclopts-based command-line client. Thin wrapper over the GraphQL API.

Cluster — A Kubernetes cluster registered with Styrmin. Cluster-level settings are inherited by every Environment and Deployment on it. One agent runs per Cluster.

Component — A logical sub-part of an application (e.g. server, task-worker) declared by the Application Driver and identified by a Kubernetes label selector. Lifecycle controls (start/stop, restart, scale) act on the Component as a unit.

Component overrides — The mechanism by which a Deployment overrides the Application Driver Version's default Component config. First-class to the deployment-config feature.

Context (GlobalContext) — The resolved configuration snapshot both the server and the agent act on. Embedded in both the Instance's context field and the StyrminDeployment CRD spec. Frozen for the lifetime of an Instance. See the Context page.

D

Deployment — The long-lived record carrying the desired state of one application in one Environment: which Application Driver Version, which Application Version, and the user's config. Editing it triggers reconciliation.

DeploymentStatus — The latest StyrminDeployment CRD .status snapshot per deployment. Distinct from the Deployment state enum.

Deployment state — The enum on Deployment.status: PENDING, DEPLOYING, RUNNING, RESTORING, FAILED, DELETING, DELETED.

Dual-entrypoint E2E — E2E tests that run twice, once driven through the SDK and once through Playwright, so both the API surface and the UI surface stay in sync.

E

E2E (test tier) — End-to-end test against a real vcluster with FluxCD, Prefect, and supporting services — no mocks.

Environment — A logical deployment target inside a Cluster — dev, staging, prod, etc. Groups Deployments that share a lifecycle and a Backup Storage Location.

F

Flow / Prefect flow — A unit of orchestration (CREATE_INSTANCE, UPDATE_INSTANCE, UPGRADE_INSTANCE, ...) submitted by the server and executed on the agent's Prefect worker.

Frontend — The React 19 + Vite single-page app that consumes the GraphQL API.

G

Generator — The server component that builds a fresh Instance from a Deployment, resolving the Context as part of the build.

GlobalContext — See Context.

GraphQL operation — A query or mutation defined in a .graphql file. Inline GraphQL strings in Python are not allowed.

H

HelmRelease — The FluxCD v2 CRD declaring desired Helm chart state. Created and updated by the operator based on the resolved Context.

Hook — A lifecycle slot at which Actions run. Currently setup (during create) and upgrade (during upgrade). Each Hook has three Modes.

I

Instance (ApplicationInstance) — A frozen snapshot of what was deployed at a point in time. Generated from a Deployment at reconciliation time. A Deployment owns at most one active instance at a time.

M

Mode — The execution phase within a Hook: pre, core, or post. Ordering across modes is strict; ordering inside a mode is unspecified.

O

Observed state — What the agent actually sees in the cluster, as opposed to declared state. Captured in PodStatus and DeploymentStatus. Drift between observed and declared is expected.

Operator — The kopf-based controller running inside the agent that watches StyrminDeployment CRDs and reconciles them into Kubernetes resources. See The operator and StyrminDeployment.

P

Parameter — A user-configurable input declared by the Application Driver. Surfaced through the GraphQL API and frontend forms.

PodStatus — Per-pod observed-state record (one row per (deployment_id, pod_name)).

Prefect — The workflow engine Styrmin uses to orchestrate Flows on the agent.

R

Reconciliation — The act of comparing the Deployment's newly generated Instance against the active Instance and planning the convergence work. The core loop of Styrmin. See Reconciliation.

Recorded intent — The flow of direction from server to agent through database rows and CRDs, not direct RPC. The server never calls the agent directly.

Remediator — The server service that maps the diff between a newly generated Instance and the active Instance to a Prefect flow.

S

SDK / styrmin-sdk — The async Python client SDK that wraps GraphQL operations as typed Pydantic models.

Server — The Styrmin backend process. Owns the database, exposes the GraphQL API. The single source of truth for what should be running.

Service (deployed service) — A network-addressable endpoint declared by the Application Driver and materialised as a Kubernetes Service object.

Status reporter — The coroutine inside the agent that polls Kubernetes and posts PodStatus / DeploymentStatus snapshots back to the server. See Status reporting.

StyrminDeployment (CRD) — Styrmin's primary Custom Resource Definition. Carries the embedded Context in spec; watched by the operator.

T

Task (database record) — A row tracking one asynchronous operation, with status, type, and the Prefect flow run id. Not to be confused with Prefect's in-process task primitive.

Tombstone / tombstoned deployment — A Deployment row whose cluster-side workload has been torn down but whose record is retained for timeline, backups, and audit.

U

Utility — A cluster-level resource (ingress controller, service discovery, ...) managed by the same builders that produce application resources. Internal to Styrmin; not user-facing in normal use.

V

vcluster — The virtual Kubernetes cluster used as the E2E test target and the local-demo runtime. Bundled with the Styrmin repo's Quickstart workflow.