Skip to content

Process Management API

Use these endpoints to list services, inspect startup state, and start, stop, or restart a managed process. Process names must match the names returned by GET /process/processes; do not infer them from configuration keys.

All paths are backend-native. When calling through port 3005, prefix REST paths with /api. Authentication is enforced by the backend.

Visual guide: See the backend control-plane and request flows.

Find the right workflow

Task Reference
Service lifecycle, status, discovery, and runtime log level This page
Check or install updates; maintain the install cache Updates and install cache
Repair, back up, compare, or restore symlinks Symlink operations
Move a supported application from SQLite to PostgreSQL PostgreSQL migration
Benchmark an InfiniDysk-backed rclone mount Rclone optimizer
Rename a legacy NzbDAV installation to InfiniDysk Namespace migration
Dev Plan and complete a service arrangement Setup
Dev Back up or restore a managed installation Installation backups
Dev Register an archive destination Backup storage
Dev Install or migrate to native InfiniDysk 2 InfiniDysk 2

Service endpoints

GET /process/processes

Returns all configured processes, including enabled status, version, repo URL, sponsorship URL, manual-update support, and the last recorded update status. The dashboard uses the latter two fields to build its update inventory without one status request per service.

Example Response:

{
  "processes": [
    {
      "name": "rclone w/ RealDebrid",
      "process_name": "rclone w/ RealDebrid",
      "enabled": true,
      "config": { "enabled": true, "...": "..." },
      "version": "1.65.1",
      "key": "rclone",
      "config_key": "rclone",
      "repo_url": "https://rclone.org",
      "sponsorship_url": "https://rclone.org/sponsor/",
      "supports_manual_update": true,
      "update_status": {
        "status": "update_available",
        "current_version": "1.65.1",
        "available_version": "1.66.0"
      }
    }
  ]
}

GET /process

Fetch details about a specific process.

Required Query Parameter:

  • process_name (string)

Example Response:

{
  "process_name": "rclone w/ RealDebrid",
  "config": { "enabled": true, "...": "..." },
  "version": "1.65.1",
  "config_key": "rclone",
  "repo_url": "https://rclone.org",
  "sponsorship_url": "https://rclone.org/sponsor/"
}

POST /process/start-service

Starts a specific process.

Request Body:

{
  "process_name": "rclone w/ RealDebrid"
}

Example Response:

{
  "status": "Service started successfully",
  "process_name": "rclone w/ RealDebrid"
}

POST /process/stop-service

Stops a running process.

Request Body:

{
  "process_name": "rclone w/ RealDebrid"
}

POST /process/restart-service

Restarts a running process.

Request Body:

{
  "process_name": "rclone w/ RealDebrid"
}

GET /process/service-reset/preview

Returns the guarded reset/removal plan when capability service_reset=true.

Required query parameter: process_name. Optional action is reset (default) or remove. The response identifies whether DUMB will reset a required template or remove a custom instance, reports default_instance_after_removal when removing the final custom instance will restore the disabled default template, lists exact eligible file targets, and reports retained/shared paths and dependency-reference warnings. Preview is read-only.

POST /process/service-reset

Stops one eligible service, cancels its DUMB schedules, writes a private full-config backup, then applies a freshly validated preview.

{
  "process_name": "Sonarr Movies",
  "action": "remove",
  "confirmation": "Sonarr Movies"
}

confirmation must exactly match process_name. reset preserves application files. remove clears only the returned DUMB-owned paths. Neither action removes mounts, symlink libraries, shared caches, PostgreSQL databases, external data stores, or other service configuration. DUMB API and DUMB Frontend are excluded.

See Service Reset and Removal for template behavior, safety boundaries, and recovery.

GET /process/service-status

Gets the current status of a process.

Pass include_health=true to include structured health and Auto-restart state. The health_status value is healthy, degraded, starting, or unhealthy. healthy remains for backward compatibility; use health_status when the difference between an application starting, degraded, and restart-worthy unhealthy state matters.

Example Response:

{
  "process_name": "InfiniDysk",
  "status": "running",
  "healthy": true,
  "health_status": "starting",
  "health_reason": "InfiniDysk reports migrating",
  "health_details": {
    "probe": "InfiniDysk backend health",
    "endpoint": "/health",
    "supported": true,
    "http_status": 503,
    "reported_status": "migrating"
  },
  "restart": {}
}

Health details expose only bounded probe metadata and component names/statuses. They do not include response bodies, credentials, or arbitrary remote URLs.

GET /process/startup-status

Returns the readiness-aware container startup lifecycle when capability startup_lifecycle is true. The response includes phase/timestamps, expected service names, each service's pending, starting, or ready state, and terminal failure summaries.

{
  "phase": "stabilizing",
  "terminal": false,
  "expected_services": ["DUMB API", "DUMB Frontend", "InfiniDysk"],
  "services": {
    "DUMB API": {"state": "ready", "reason": null},
    "DUMB Frontend": {"state": "ready", "reason": null},
    "InfiniDysk": {
      "state": "starting",
      "reason": "InfiniDysk reports migrating",
      "health_status": "starting",
      "health_details": {
        "probe": "InfiniDysk backend health",
        "endpoint": "/health",
        "http_status": 503,
        "reported_status": "migrating"
      }
    }
  },
  "failures": {}
}

Runtime API logging

These endpoints are available when /process/capabilities advertises runtime_api_log_level.

GET /process/runtime-log-level

Returns the configured and effective DUMB API logging levels and whether a temporary DEBUG override is active.

{
  "configured_level": "INFO",
  "configured_uvicorn_level": "INFO",
  "effective_level": "DEBUG",
  "debug_enabled": true,
  "override_active": true,
  "temporary": true,
  "resets_on_restart": true
}

POST /process/runtime-log-level

Enables or disables a temporary DEBUG override without restarting the DUMB container or API. Enabling changes the running shared DUMB logger and Uvicorn loggers immediately. Disabling restores their configured levels.

{
  "debug_enabled": true
}

The override is intentionally not written to dumb_config.json and is cleared by a container restart. It does not modify managed-service logging settings. DEBUG output can be substantially larger and may contain additional operational details; DUMB's normal log redaction remains active.

POST /process/start-core-service

Starts one or more core services and all required dependencies. This is used during onboarding.

The core_services field can be a single object or an array. The name can be the config key (e.g., riven_backend) or a display name (e.g., Riven).

Request Body Examples:

{
  "core_services": {
    "name": "riven_backend",
    "debrid_service": "RealDebrid",
    "debrid_key": "abc123",
    "service_options": {}
  },
  "optional_services": ["zilean", "pgadmin", "riven_frontend"]
}
{
  "core_services": {
    "name": "decypharr",
    "debrid_service": "RealDebrid",
    "debrid_key": "abc123",
    "service_options": {
      "decypharr": {
        "mount_type": "dfs",
        "mount_path": "/mnt/debrid/decypharr"
      }
    }
  },
  "optional_services": []
}
{
  "core_services": {
    "name": "cli_debrid",
    "debrid_service": "RealDebrid",
    "debrid_key": "abc123",
    "service_options": {
      "phalanx_db": { "enabled": true }
    }
  },
  "optional_services": ["zilean"]
}

Example Response:

{
  "results": [
    {"service": "riven_backend", "status": "started"},
    {"service": "decypharr", "status": "started"}
  ],
  "errors": []
}

Notes

  • Dependencies like Zurg or rclone are created using templates and attached to the calling core service.
  • Optional services such as pgadmin, zilean, bazarr, pulsarr, maintainerr, mediastorm, traefik_proxy_admin, or cloudflared are started only if included.
  • debrid_key is injected into Zurg or Decypharr as needed.
  • service_options can override config values such as log_level, port, or enabled.
  • Any startup errors appear in the errors list.

GET /process/core-services

Returns the available core services, their dependencies, and default service options (used by onboarding).

GET /process/dependency-graph

Returns backend-resolved dependency relationships for a specific process, including:

  • static core dependencies (with instance-scoped matching for rclone/zurg)
  • inferred links from core_service/core_services
  • inferred links from wait_for_url/wait_for_dir entries (including localhost port matching)
  • inferred links from wait_for_mounts by matching required mount paths to provider service mount points
  • rclone provider links (zurg_enabled, decypharr_enabled, key_type=infinidysk) to reflect WebDAV provider dependencies directly
  • non-core hard dependency map for service-specific startup requirements (for example riven_frontend -> riven_backend, zilean -> postgres, pgadmin -> postgres)
  • conditional startup dependencies from the backend startup ordering logic -- config-aware dependencies like tautulli -> plex (when plex is enabled), bazarr -> sonarr/radarr (when those Arrs are enabled), prowlarr -> sonarr/radarr (when those Arrs are enabled), neutarr -> sonarr (when Sonarr has use_neutarr enabled), etc. For instance-scoped services (rclone, zurg), conditional deps are filtered per-instance so that only the specific associated instances are shown (for example, a rclone instance with zurg_enabled only shows its own zurg instance, not zurg instances belonging to other rclone configurations)
  • documented integration links (soft linkage, scope=all only) -- for example seerr -> sonarr/radarr request routing
  • per-process status state used by the frontend dependency graph panel
  • dependency_truth_table describing the 12 dependency signal types and whether each is treated as hard dependency vs linkage
  • signals array on each row/edge identifying which detection signals established the relationship (for example ["core_service_map"], ["rclone_provider_zurg", "conditional_startup_map"])

Required Query Parameter:

  • process_name (string)

Optional Query Parameter:

  • scope (string): runtime (default) or all
    • runtime returns runtime/configured hard dependencies
    • all additionally includes soft linkage edges (for example optional Zilean integrations)

The response includes parallel_groups metadata describing concurrent prerequisite/dependent stages around the selected service.

Example Response (shape):

{
  "process_name": "Seerr Main",
  "config_key": "seerr",
  "context": { "mode": "core", "key": "seerr", "core": { "key": "seerr", "name": "Seerr", "dependencies": [] } },
  "scope": "runtime",
  "startup_order": [{ "key": "seerr", "label": "Seerr", "state": "running" }],
  "dependency_rows": [],
  "dependent_rows": [],
  "linked_outgoing_rows": [{ "process_name": "Plex Media Server", "key": "plex", "label": "Plex Media Server", "state": "running" }],
  "linked_incoming_rows": [],
  "nodes": [{ "id": "Seerr Main", "process_name": "Seerr Main", "key": "seerr", "label": "Seerr Main", "state": "running" }],
  "edges": [{ "source": "Seerr Main", "target": "Plex Media Server", "signals": ["wait_for_url"], "strength": "hard_runtime" }],
  "parallel_groups": [{ "id": "pre_core", "label": "Parallel prerequisites", "type": "parallel", "members": ["PostgreSQL", "Rclone w/ Riven"] }],
  "updated_at": "2026-02-11T18:45:00Z"
}

GET /process/optional-services

Returns optional services. You can pass core_service and optional_services query params to tailor the list.

GET /process/mediastorm-initial-admin-password

Returns mediastorm's active bootstrap credential while its first-login password file exists:

{
  "available": true,
  "username": "admin",
  "password": "admin",
  "credential_kind": "default"
}

After the administrator changes the password, mediastorm deletes the file and the endpoint returns:

{
  "available": false,
  "username": "admin",
  "password": null,
  "credential_kind": null
}

credential_kind is default when the file contains mediastorm's current public admin password and installation_specific for an older generated password or an explicit STRMR_INITIAL_ADMIN_PASSWORD value.

On current mediastorm builds, clients using the public default must provide a replacement newPassword in the same login request. A login containing only admin / admin returns HTTP 428 with code password_change_required; the replacement is committed before the first session is created.

The endpoint uses the existing DUMB authentication dependency, accepts no caller-supplied path, and reads only the fixed mediastorm cache credential. It refuses symlinks, non-regular files, oversized values, and multiline values. Responses include Cache-Control: no-store, private and Pragma: no-cache; clients must not persist the password.

DUMB checks both initial_admin_password and initial_admin_password.txt through mediastorm's configured persistent cache directory. Clients should gate this endpoint on the mediastorm_initial_admin_password capability.

Capabilities

GET /process/capabilities

Read capabilities before offering an optional workflow. The response is the authority for that running backend; the docs channel or the dashboard version alone cannot establish support.

curl -H "Authorization: Bearer $DUMB_ACCESS_TOKEN" \
  http://localhost:3005/api/process/capabilities

Common capability groups:

Fields Controls
manual_update_check, configured_source_install, commit_sha_pinning Manual update checks and saved source targets
auto_update_start_time, auto_update_mode Schedule anchors and report-only schedules
startup_lifecycle Readiness-aware GET /process/startup-status
runtime_api_log_level Temporary DEBUG control for the running API
symlink_* Individual repair, manifest, job, restore, and scheduling operations
postgres_migration*, postgres_migration_service_keys Guarded database migration and the supported service list
metrics_history_storage, metrics_history_hot_activation History providers and live PostgreSQL activation
metrics_filesystem_selection, metrics_network_interface_selection Selected filesystem and interface metrics
infinidysk_migration* Legacy identity and namespace migration
rclone_optimizer_infinidysk InfiniDysk generation 1 rclone optimizer
mediastorm_initial_admin_password No-store bootstrap credential readback
Dev setup_arrangements Retained service setup plans and action runs
Dev setup_infinidysk_runtime_choices, setup_infinidysk_runtime_protocol Native runtime selection handoff; protocol 1
Dev installation_backup, installation_backup_protocol Installation backup/restore; protocol 1
Dev installation_backup_cancellation, installation_backup_phase_progress Job-specific cancellation and measured phase progress
Dev backup_storage_selection, backup_storage_protocol Registered archive destinations; protocol 1
Dev backup_storage_shared_permissions Explicit shared-permissions consent per storage profile
Dev infinidysk_v2, infinidysk_v2_migration, infinidysk_v2_migration_protocol Native generation 2 workflow; protocol 1
Dev infinidysk_v2_postgres Native generation 2 PostgreSQL conversion
Dev infinidysk_v2_async_preview Background migration preview with progress, cancellation and result retrieval
Dev infinidysk_v2_consumer_recovery, infinidysk_v2_consumer_recovery_protocol Retained target/consumer recovery; protocol 1
Dev infinidysk_v2_cleanup, infinidysk_v2_cleanup_protocol Guarded v1 retirement; protocol 1
Dev infinidysk_v2_low_space_migration, infinidysk_v2_external_blob_staging Explicit capacity-saving migration choices

Require the exact protocol advertised by the relevant workflow. Treat an absent flag as unsupported. Legacy arr_postgres_migration*, nzbdav_install_info, and rclone_optimizer_nzbdav aliases exist for older clients; new clients should use canonical names.

Generation 2 also requires valid api_version SemVer at least 2.24.0. Its runtime, mount backend, and database provider are separate choices. infinidysk_v2_rclone_optimizer is currently false, including when a generation 2 installation retains rclone.

Previously on this page

These section links remain available for existing bookmarks:

Setup plans and retained runs

See Setup API.

Install cache

See install cache endpoints.

Updates and scheduling

See Updates and install cache API.

Installation backup and restore

See Installation backups API.

Registered backup storage

See Backup storage API.

See Symlink operations API.

SQLite-to-PostgreSQL Migration

See PostgreSQL migration API.

Rclone streaming optimizer

See Rclone optimizer API.

InfiniDysk 2 workflows

See InfiniDysk 2 API.

InfiniDysk migration

See Namespace migration API.