# Architecture Guide

## Overview

Certctl is a certificate management platform with a **decoupled control-plane and agent architecture**. The control plane orchestrates certificate issuance and renewal, while agents deployed across your infrastructure handle key generation, certificate deployment, and local validation — private keys never leave the infrastructure they were generated on.

New to certificates? Read the [Concepts Guide](concepts.md) first.

### Design Principles

1. **Private Key Isolation** — Agents generate ECDSA P-256 keys locally and submit CSRs only. Private keys never touch the control plane. Server-side keygen available via `CERTCTL_KEYGEN_MODE=server` for demo only.
2. **GUI as Primary Interface** — The web dashboard is the operational control plane, not a secondary viewer. Every backend feature ships with its corresponding GUI surface.
3. **Decoupled Operations** — Agents operate autonomously; the control plane coordinates but doesn't block agent function
4. **Audit-First** — Complete traceability of all issuance, deployment, and rotation events
5. **Connector Architecture** — Pluggable issuers, targets, and notifiers for extensibility
6. **Self-Hosted** — No cloud lock-in; run with Docker Compose, Kubernetes, or bare metal

## System Components

```mermaid
flowchart TB
    subgraph "Control Plane"
        API["REST API\n(Go net/http, :8443)"]
        SVC["Service Layer"]
        REPO["Repository Layer\n(database/sql + lib/pq)"]
        SCHED["Background Scheduler\n4 loops"]
        DASH["Web Dashboard\n(React SPA)"]
    end

    subgraph "Data Store"
        PG[("PostgreSQL 16\n14 tables\nTEXT primary keys")]
    end

    subgraph "Agent Fleet"
        A1["Agent: nginx-prod\n(heartbeat + work poll)"]
        A2["Agent: f5-prod"]
        A3["Agent: iis-prod"]
    end

    subgraph "Issuer Backends"
        CA1["Local CA\n(crypto/x509)"]
        CA2["ACME\n(Let's Encrypt)"]
        CA3["Vault PKI\n(future)"]
    end

    subgraph "Target Systems"
        T1["NGINX\n(SSH + reload)"]
        T2["F5 BIG-IP\n(REST API)"]
        T3["IIS\n(WinRM)"]
    end

    DASH --> API
    API --> SVC
    SVC --> REPO
    REPO --> PG
    SCHED --> SVC
    SVC -->|"Issue/Renew"| CA1 & CA2 & CA3

    A1 & A2 & A3 -->|"CSR + Heartbeat"| API
    API -->|"Cert + Chain\n(NO private key)"| A1 & A2 & A3

    A1 -->|"Deploy"| T1
    A2 -->|"Deploy"| T2
    A3 -->|"Deploy"| T3
```

### Control Plane (Server)

The control plane is a Go HTTP server backed by PostgreSQL. It manages state (certificates, agents, targets, issuers, policies), orchestrates issuance by coordinating with CAs through issuer connectors, tracks jobs for certificate issuance/renewal/deployment workflows, maintains an immutable audit trail, and dispatches work via a background scheduler.

The server exposes a REST API under `/api/v1/` and optionally serves the web dashboard as static files from the `web/` directory.

**Key internals**: The server uses Go 1.22's `net/http` stdlib routing (no external router framework), structured logging via `slog`, and a handler → service → repository layered architecture. Handlers define their own service interfaces for clean dependency inversion.

### Agents

Lightweight Go processes that run on or near your infrastructure. Agents generate ECDSA P-256 private keys locally, create CSRs, and submit them to the control plane for signing — private keys never leave agent infrastructure. Agents also handle certificate deployment to target systems (NGINX, F5, IIS) and report job status. They communicate with the control plane via HTTP and authenticate with API keys.

The agent runs two background loops: a heartbeat (every 60 seconds) to signal it's alive, and a work poll (every 30 seconds) to check for actionable jobs via `GET /api/v1/agents/{id}/work`. Jobs may be `AwaitingCSR` (agent needs to generate key + submit CSR) or `Deployment` (agent needs to deploy a certificate). Private keys are stored in `CERTCTL_KEY_DIR` (default `/var/lib/certctl/keys`) with 0600 permissions.

### Web Dashboard

The web dashboard is the primary operational interface for certctl. It is built with Vite + React + TypeScript and uses TanStack Query for server state management (caching, background refetching, optimistic updates).

**Current views**: certificate inventory (list with "New Certificate" creation modal + detail with version history, deploy, archive, and trigger renewal actions), agent fleet (health indicators from heartbeat), job queue (status, retry, cancel), notification inbox (threshold alert grouping, mark-as-read), audit trail (time range and actor/action filters), policy management (rules with enable/disable toggle + delete + violations), issuers (list with test connection + delete), targets (list with delete), and a summary dashboard.

The dashboard includes a **demo mode** that activates when the API is unreachable — it renders realistic mock data for screenshots and offline presentations.

**Tech decisions**:
- Vite for fast builds and HMR during development
- TanStack Query over manual fetch/useEffect for automatic cache invalidation and refetching
- Dark theme default (ops teams live in dark mode)
- SSE/WebSocket planned for real-time job status updates (V2.0)

### PostgreSQL Database

All state is stored in PostgreSQL 16. The schema uses TEXT primary keys (not UUIDs) with human-readable prefixed IDs like `mc-api-prod`, `t-platform`, `o-alice`.

```mermaid
erDiagram
    teams ||--o{ owners : "has members"
    teams ||--o{ managed_certificates : "owns"
    owners ||--o{ managed_certificates : "responsible for"
    issuers ||--o{ managed_certificates : "signs"
    renewal_policies ||--o{ managed_certificates : "governs"
    managed_certificates ||--o{ certificate_versions : "has versions"
    managed_certificates ||--o{ certificate_target_mappings : "deployed to"
    deployment_targets ||--o{ certificate_target_mappings : "receives"
    agents ||--o{ deployment_targets : "manages"
    managed_certificates ||--o{ jobs : "triggers"
    policy_rules ||--o{ policy_violations : "produces"
    managed_certificates ||--o{ policy_violations : "violates"
    managed_certificates ||--o{ audit_events : "logged in"
    managed_certificates ||--o{ notification_events : "generates"

    teams {
        text id PK
        text name
        text description
    }
    owners {
        text id PK
        text name
        text email
        text team_id FK
    }
    managed_certificates {
        text id PK
        text name
        text common_name
        text[] sans
        text environment
        text owner_id FK
        text team_id FK
        text issuer_id FK
        text renewal_policy_id FK
        text status
        timestamp expires_at
        jsonb tags
    }
    certificate_versions {
        text id PK
        text certificate_id FK
        text serial_number
        text fingerprint_sha256
        text pem_chain
        text csr_pem
    }
    agents {
        text id PK
        text name
        text hostname
        text status
        text api_key_hash
    }
    deployment_targets {
        text id PK
        text name
        text type
        text agent_id FK
        jsonb config
    }
    issuers {
        text id PK
        text name
        text type
        jsonb config
        boolean enabled
    }
    jobs {
        text id PK
        text type
        text certificate_id FK
        text target_id FK
        text status
        int attempts
    }
    policy_rules {
        text id PK
        text name
        text type
        jsonb config
        boolean enabled
    }
    policy_violations {
        text id PK
        text certificate_id FK
        text rule_id FK
        text message
        text severity
    }
    audit_events {
        text id PK
        text actor
        text actor_type
        text action
        text resource_type
        text resource_id
        jsonb details
    }
    notification_events {
        text id PK
        text type
        text certificate_id FK
        text channel
        text recipient
        text status
    }
```

Migrations are idempotent (`IF NOT EXISTS` on all CREATE statements, `ON CONFLICT (id) DO NOTHING` on all seed data) so they're safe to run multiple times — important for Docker Compose where both initdb and the server may run the same SQL.

## Data Flow: Certificate Lifecycle

### 1. Create Managed Certificate

```mermaid
sequenceDiagram
    participant U as User / API Client
    participant API as REST API
    participant SVC as CertificateService
    participant DB as PostgreSQL
    participant AUD as AuditService

    U->>API: POST /api/v1/certificates<br/>{name, common_name, sans, ...}
    API->>SVC: Create(ctx, certificate)
    SVC->>SVC: Validate required fields
    SVC->>DB: INSERT INTO managed_certificates
    SVC->>AUD: Create(audit_event: certificate_created)
    AUD->>DB: INSERT INTO audit_events
    SVC-->>API: ManagedCertificate
    API-->>U: 201 Created + JSON body
```

### 2. Certificate Issuance

#### Agent-Side Key Generation (Default)

In the default `agent` keygen mode (`CERTCTL_KEYGEN_MODE=agent`), the control plane never touches private keys. When a renewal or issuance job is created, it enters `AwaitingCSR` state. The agent picks it up, generates an ECDSA P-256 key pair locally, and submits only the CSR (public key).

```mermaid
sequenceDiagram
    participant S as Scheduler
    participant SVC as RenewalService
    participant DB as PostgreSQL
    participant A as Agent
    participant API as Control Plane API
    participant ISS as Issuer Connector

    S->>SVC: ProcessRenewalJob(job)
    SVC->>DB: UPDATE job SET status='AwaitingCSR'
    SVC->>DB: UPDATE cert SET status='RenewalInProgress'

    A->>API: GET /agents/{id}/work
    API-->>A: [{id, type:"Renewal", status:"AwaitingCSR", common_name, sans}]

    A->>A: Generate ECDSA P-256 key pair
    A->>A: Store key to CERTCTL_KEY_DIR/certId.key (0600)
    A->>A: Create CSR with CN + SANs

    A->>API: POST /agents/{id}/csr<br/>{csr_pem, certificate_id}
    API->>SVC: CompleteAgentCSRRenewal(job, cert, csrPEM)
    SVC->>ISS: RenewCertificate(CN, SANs, csrPEM)
    ISS-->>SVC: IssuanceResult{cert_pem, chain_pem, serial}
    SVC->>DB: INSERT INTO certificate_versions (PEM chain + CSR only)
    SVC->>DB: UPDATE cert SET status='Active', expires_at
    SVC->>DB: CREATE deployment jobs for targets

    Note over A: Agent deploys using locally-held private key
```

#### Server-Side Key Generation (Demo Only)

Set `CERTCTL_KEYGEN_MODE=server` for development/demo with Local CA. The control plane generates RSA-2048 keys server-side. A log warning is emitted at startup.

```mermaid
sequenceDiagram
    participant U as User / Scheduler
    participant SVC as RenewalService
    participant ISS as IssuerConnector
    participant DB as PostgreSQL

    U->>SVC: ProcessRenewalJob(job)
    SVC->>SVC: Generate RSA-2048 key pair (server-side)
    SVC->>SVC: Create CSR with CN + SANs
    SVC->>ISS: RenewCertificate(CN, SANs, csrPEM)
    ISS-->>SVC: IssuanceResult{cert_pem, chain_pem, serial}
    SVC->>DB: INSERT INTO certificate_versions (PEM + private key)
    SVC->>DB: UPDATE cert SET status='Active'
    SVC->>DB: CREATE deployment jobs

    Note over SVC: WARNING: Private keys touch control plane
```

### 3. Deploy Certificate to Target

The agent deploys certificates using target connectors. Each connector knows how to push certificates to a specific system:

- **NGINX**: Writes cert/chain files to disk, validates config with `nginx -t`, reloads with `nginx -s reload` or `systemctl reload nginx`
- **F5 BIG-IP**: Calls the F5 REST API to upload certificate and update virtual server bindings
- **IIS**: Uses WinRM to import the certificate into the Windows certificate store and bind it to an IIS site

The agent handles both the certificate (public) and the private key (read from local key store at `CERTCTL_KEY_DIR`). The control plane never sees the private key.

### 4. Automatic Renewal

The control plane runs a scheduler with four background loops:

```mermaid
flowchart LR
    subgraph "Scheduler (Background Goroutines)"
        R["Renewal Checker\n⏱ every 1h"]
        J["Job Processor\n⏱ every 30s"]
        H["Agent Health\n⏱ every 2m"]
        N["Notification Processor\n⏱ every 1m"]
    end

    R -->|"Find expiring certs\nCreate renewal jobs"| DB[("PostgreSQL")]
    J -->|"Process pending jobs\nCoordinate issuance"| DB
    H -->|"Check heartbeat staleness\nMark agents offline"| DB
    N -->|"Send pending notifications\nEmail / Webhook"| DB
```

| Loop | Interval | Purpose |
|------|----------|---------|
| Renewal checker | 1 hour | Finds certificates approaching expiry, creates renewal jobs |
| Job processor | 30 seconds | Processes pending jobs (issuance, renewal, deployment) |
| Agent health check | 2 minutes | Marks agents as offline if heartbeat is stale |
| Notification processor | 1 minute | Sends pending notifications via configured channels |

When the renewal checker finds a certificate within its renewal window, it performs two tasks: threshold-based alerting and renewal job creation.

**Threshold-Based Expiration Alerting**: Each renewal policy defines configurable alert thresholds (default: 30, 14, 7, 0 days before expiry). For each certificate approaching expiry, the scheduler checks which thresholds have been crossed and sends deduplicated notifications. A certificate that crosses the 14-day threshold only gets one 14-day alert, even though the renewal checker runs every hour. Deduplication is tracked via threshold tags embedded in the notification message and queried with the `MessageLike` filter. Certificates are also transitioned to `Expiring` status when they enter the alert window and `Expired` when they hit 0 days.

**Renewal Job Creation**: If the certificate's issuer has a registered connector, the scheduler creates a renewal job. The job processor picks it up, coordinates with the issuer, and triggers deployment. All steps are logged in the audit trail and generate notifications.

## Connector Architecture

Certctl uses connector interfaces for extensibility. Each connector type has a standard interface that implementations must satisfy.

```mermaid
flowchart TB
    subgraph "Issuer Connectors"
        direction TB
        II["IssuerConnector Interface\nIssueCertificate() | RenewCertificate()\nRevokeCertificate() | GetOrderStatus()"]
        II --> LC["Local CA"]
        II --> ACME["ACME v2"]
        II --> VP["Vault PKI (future)"]
    end

    subgraph "Target Connectors"
        direction TB
        TI["TargetConnector Interface\nDeployCertificate()\nValidateDeployment()"]
        TI --> NG["NGINX"]
        TI --> F5["F5 BIG-IP"]
        TI --> IIS["IIS"]
    end

    subgraph "Notifier Connectors"
        direction TB
        NI["NotifierConnector Interface\nSendAlert() | SendEvent()"]
        NI --> EM["Email (SMTP)"]
        NI --> WH["Webhook (HTTP)"]
        NI --> SL["Slack (future)"]
    end
```

### IssuerConnectorAdapter (Dependency Inversion)

The service layer defines its own `IssuerConnector` interface (`internal/service/renewal.go`) while the connector layer has its own `issuer.Connector` interface (`internal/connector/issuer/interface.go`). The `IssuerConnectorAdapter` (`internal/service/issuer_adapter.go`) bridges the two, translating between their request/response types. This maintains clean dependency inversion — the service package never imports the connector package directly.

```mermaid
flowchart LR
    SVC["Service Layer<br/>service.IssuerConnector"] --> ADAPT["IssuerConnectorAdapter<br/>(bridges interfaces)"]
    ADAPT --> CONN["Connector Layer<br/>issuer.Connector"]
    CONN --> LC["Local CA"]
    CONN --> ACME["ACME v2"]
```

Registration happens in `cmd/server/main.go`:
```go
localCA := local.New(nil, logger)
issuerRegistry := map[string]service.IssuerConnector{
    "iss-local": service.NewIssuerConnectorAdapter(localCA),
}
```

### Issuer Connector

Handles certificate issuance from CAs.

```go
type Connector interface {
    ValidateConfig(ctx context.Context, config json.RawMessage) error
    IssueCertificate(ctx context.Context, request IssuanceRequest) (*IssuanceResult, error)
    RenewCertificate(ctx context.Context, request RenewalRequest) (*IssuanceResult, error)
    RevokeCertificate(ctx context.Context, request RevocationRequest) error
    GetOrderStatus(ctx context.Context, orderID string) (*OrderStatus, error)
}
```

Built-in issuers: **Local CA** (self-signed, in-memory CA for development/demos using `crypto/x509`) and **ACME v2** (fully implemented with HTTP-01 challenge solving, compatible with Let's Encrypt, Sectigo, and any ACME-compliant CA). The ACME connector uses `golang.org/x/crypto/acme`, generates an ECDSA P-256 account key, handles account registration with ToS acceptance, order creation, HTTP-01 challenge solving via a built-in temporary HTTP server, order finalization, and DER-to-PEM chain conversion. Configure via `CERTCTL_ACME_DIRECTORY_URL` and `CERTCTL_ACME_EMAIL`.

### Target Connector

Deploys certificates to infrastructure. The `DeploymentRequest` includes `KeyPEM` because agents generate and hold private keys locally — the key is passed from the agent's local key store into the target connector, never from the control plane.

```go
type Connector interface {
    ValidateConfig(ctx context.Context, config json.RawMessage) error
    DeployCertificate(ctx context.Context, request DeploymentRequest) (*DeploymentResult, error)
    ValidateDeployment(ctx context.Context, request ValidationRequest) (*ValidationResult, error)
}
```

The `DeploymentRequest` struct carries the full material needed by the target system: the signed certificate, the CA chain, the agent-generated private key, target-specific configuration, and arbitrary metadata. The key field is populated by the agent from its local key store (`CERTCTL_KEY_DIR`) — it never originates from the control plane.

Built-in targets: **NGINX** (writes cert/chain/key files, validates with `nginx -t`, reloads), **F5 BIG-IP** (REST API upload + virtual server binding), **IIS** (WinRM PFX import + site binding).

### Notifier Connector

Sends alerts about certificate lifecycle events.

```go
type Connector interface {
    ValidateConfig(ctx context.Context, config json.RawMessage) error
    SendAlert(ctx context.Context, alert Alert) error
    SendEvent(ctx context.Context, event Event) error
}
```

Built-in notifiers: **Email** (SMTP) and **Webhook** (HTTP POST).

See the [Connector Development Guide](connectors.md) for details on building custom connectors.

## Security Model

### Private Key Management

```mermaid
flowchart LR
    subgraph "Agent (Your Infrastructure)"
        GEN["1. GENERATE\ncrypto/ecdsa P-256"]
        STORE["2. STORE\nFile perms 0600"]
        USE["3. USE\nCSR gen + deployment"]
        ROT["4. ROTATE\nDelete old after renewal"]
    end

    subgraph "Control Plane (certctl-server)"
        CP["Only sees:\n• Certificates (public)\n• Chains (public)\n• CSRs (public key only)"]
    end

    GEN --> STORE --> USE --> ROT
    USE -.->|"CSR (public key only)"| CP
    CP -.->|"Signed cert + chain"| USE

    style CP fill:#fee,stroke:#c33
    style GEN fill:#efe,stroke:#3c3
    style STORE fill:#efe,stroke:#3c3
    style USE fill:#efe,stroke:#3c3
    style ROT fill:#efe,stroke:#3c3
```

**Agent keygen mode (default, `CERTCTL_KEYGEN_MODE=agent`):** Private keys follow a strict lifecycle on agents:

1. **Generated on the agent** — ECDSA P-256, never sent to the control plane
2. **Stored on the agent** — `CERTCTL_KEY_DIR` with file permissions 0600
3. **Used by the agent** — for deployment to targets (via `DeploymentRequest.KeyPEM`)
4. **Rotated by the agent** — old keys overwritten after successful renewal

The control plane only handles public material: certificates, chains, and CSRs.

**Server keygen mode (`CERTCTL_KEYGEN_MODE=server`, demo only):** The control plane generates RSA-2048 keys server-side within `processRenewalServerKeygen`. Private keys are stored in `certificate_versions.csr_pem`. A log warning is emitted at startup. Use only for Local CA development/demo.

### Authentication

- **API clients → Server**: API key in `Authorization: Bearer` header, or `none` for demo mode
- **Agent → Server**: API key registered at agent creation, included in all requests
- **Server → Issuers**: ACME account key, or connector-specific credentials
- **Agent → Targets**: SSH keys, API tokens, WinRM credentials (stored locally on agent)

### Audit Trail

Every action is recorded as an immutable audit event:

```json
{
  "id": "audit-001",
  "actor": "o-alice",
  "actor_type": "User",
  "action": "certificate_created",
  "resource_type": "certificate",
  "resource_id": "mc-api-prod",
  "details": {"environment": "production"},
  "timestamp": "2026-03-14T10:30:00Z"
}
```

Audit events cannot be modified or deleted. They support filtering by actor, action, resource type, resource ID, and time range.

## API Design

All endpoints are under `/api/v1/` and follow consistent patterns:

- **List**: `GET /api/v1/{resources}` — returns `{data: [...], total, page, per_page}`
- **Get**: `GET /api/v1/{resources}/{id}` — returns the resource
- **Create**: `POST /api/v1/{resources}` — returns the created resource with `201`
- **Update**: `PUT /api/v1/{resources}/{id}` — returns the updated resource
- **Delete**: `DELETE /api/v1/{resources}/{id}` — returns `204` (soft delete/archive)
- **Actions**: `POST /api/v1/{resources}/{id}/{action}` — returns `202` for async operations

Resources: certificates, issuers, targets, agents, jobs, policies, teams, owners, audit, notifications.

Health checks live outside the API prefix: `GET /health` and `GET /ready`.

## Deployment Topologies

### Docker Compose (Development / Small Deployments)

```mermaid
flowchart TB
    subgraph "Docker Network (certctl-network)"
        SERVER["certctl-server\n:8443\nAPI + Dashboard"]
        PG[("PostgreSQL\n:5432\nSchema + Seed Data")]
        AGENT["certctl-agent\nHeartbeat + Work Poll"]
    end

    USER["Browser / curl"] -->|"HTTP :8443"| SERVER
    SERVER -->|"SQL"| PG
    AGENT -->|"HTTP (internal)"| SERVER
```

### Production (Kubernetes)

```mermaid
flowchart TB
    subgraph "Kubernetes Cluster"
        subgraph "Control Plane"
            DEP["Deployment\ncertctl-server\nreplicas: 2+"]
            CM["ConfigMap\nIssuer/target configs"]
            SEC["Secret\nAPI keys, ACME creds"]
        end

        subgraph "Data"
            SS[("StatefulSet\nPostgreSQL\nprimary + replica")]
        end

        subgraph "Agent Fleet"
            DS["DaemonSet\ncertctl-agent\n(infra nodes)"]
        end
    end

    ING["Ingress\n+ TLS termination"] --> DEP
    DEP --> SS
    DEP --> CM & SEC
    DS --> DEP
```

For production, you would also add an ingress controller, TLS termination for the certctl API itself, and external PostgreSQL (RDS, Cloud SQL, etc.).

## Testing Strategy

certctl uses a layered testing approach aligned with the handler → service → repository architecture, with 220+ tests across five layers (service, handler, integration, connector, and frontend). The goal is high-confidence regression prevention at the service and handler layers, where the most complex business logic lives, combined with integration tests that exercise the full request path from HTTP to database.

**Service layer unit tests** (`internal/service/*_test.go`) — 74 test functions across 7 files with mock repositories. These test all business logic in isolation: certificate CRUD with validation, agent lifecycle (registration, heartbeat, CSR submission with both keygen modes), job state machine (creation, processing, cancellation, retry logic), policy evaluation (all 4 rule types, violation creation), renewal and issuance flow (server-side and agent-side keygen paths), and notification deduplication (threshold tag matching, channel routing). Mock repositories are simple structs with function fields, avoiding heavy mocking frameworks — this keeps tests readable and avoids coupling to mock library APIs.

**Handler layer tests** (`internal/api/handler/*_test.go`) — 119 test functions across 7 files using Go's `httptest` package. Every handler file has a corresponding test file: certificates (20 tests), agents (20 tests), jobs (14 tests), notifications (11 tests), policies (15 tests), issuers (15 tests), and targets (14 tests). Each test file follows the same pattern: a mock service struct with function fields, `httptest.NewRecorder` for capturing responses, and a shared `contextWithRequestID()` helper. Tests cover the happy path, input validation (missing fields, invalid JSON, empty IDs), error propagation from the service layer, method-not-allowed responses, and pagination parameters.

**Integration tests** (`internal/integration/`) — Two test files exercising the full stack from HTTP request through router, handler, service, and postgres repository layers. `lifecycle_test.go` has 11 subtests covering the complete certificate lifecycle: team/owner creation, certificate creation, issuer verification, renewal trigger, job verification, agent registration, CSR submission, deployment, and status reporting. `negative_test.go` has 12 subtests covering error paths: nonexistent resource lookups (404s), invalid request bodies (malformed JSON, missing required fields), invalid CSR submission, heartbeat for nonexistent agents, wrong HTTP methods on list endpoints, empty list responses, renewal on nonexistent certificates, and expired certificate lifecycle. Both use a shared `setupTestServer()` that builds a fully-wired server with real postgres repositories and the Local CA issuer connector.

**Frontend tests** (`web/src/api/client.test.ts`, `web/src/api/utils.test.ts`) — 53 Vitest tests covering the API client and utility functions. The API client tests mock `globalThis.fetch` and verify all endpoint functions (certificates, agents, jobs, policies, issuers, targets, notifications, audit, health) send correct HTTP methods, URLs, headers, and request bodies. They also test API key management (store/retrieve/clear), auth header propagation, 401 event dispatching, and error handling (server messages, error fields, status text fallback). The utility tests use `vi.useFakeTimers()` for deterministic date testing and cover `formatDate`, `formatDateTime`, `timeAgo`, `daysUntil`, and `expiryColor`. The test environment uses jsdom with `@testing-library/jest-dom` matchers.

**CI pipeline** (`.github/workflows/ci.yml`) — Two parallel jobs: Go (build, vet, test with coverage, coverage threshold enforcement) and Frontend (TypeScript type check, Vitest test suite, Vite production build). The Go job runs all tests with `-coverprofile`, then enforces coverage thresholds: service layer must be at least 30% (current: ~34%) and handler layer must be at least 50% (current: ~61%). These thresholds act as regression floors — they can only go up. The service layer threshold is deliberately lower because much of the service code depends on postgres repositories and external connectors that require real infrastructure to test meaningfully. Connector tests are included via `./internal/connector/issuer/local/...` (the Local CA package, which has unit tests for certificate signing logic). The Frontend job runs `npx vitest run` between the TypeScript check and production build steps.

**What's not tested and why:** Postgres repository implementations (`internal/repository/postgres/`) require a real database and are tested only through integration tests, not unit tests. Target connectors (NGINX, F5, IIS) depend on real infrastructure or complex mocks. Scheduler loops are time-dependent and tested manually during development. The ACME connector requires a real ACME server (tested manually against Let's Encrypt staging). These are all candidates for future expansion as the test infrastructure matures.

## What's Next

- [Quick Start](quickstart.md) — Get certctl running locally
- [Advanced Demo](demo-advanced.md) — Issue a certificate end-to-end
- [Connector Guide](connectors.md) — Build custom connectors