Trusted Paid ASA Server Manager 3.0

Shadow Server Manager​

Shadow Server Manager runs your whole ARK: Survival Ascended fleet from one browser tab. Instead of RDP-ing into every box to babysit start.bats, chase mod updates, and hand-edit config files, you sign in with Discord and drive every machine, cluster, server, plugin, mod, backup, wipe and update from a single dashboard.

It's three pieces working together: a Host (the brain and the web UI), a lightweight Client agent on each machine that runs the actual ARK servers, and a small in-game Plugin that reports what's really happening inside each server. You only ever touch the Host - the rest just does its job in the background.

How the pieces fit​

• Host - the central console + Discord bot. Lives on one always-on machine (a small VPS is plenty). This is where you and your admins actually work.
• Client - a per-machine agent that installs, updates, starts, stops and backs up the ARK servers on that box. One per Windows host that runs servers.
• Plugin - a native AsaApi plugin that runs inside each ARK server, feeding live status back and handling in-game countdown broadcasts.

The Dashboard​

• Live overview - one landing page with fleet-wide KPI tiles (players online, servers online / starting, machines online), active alerts, per-cluster status cards and a host-health table, all auto-refreshing on their own
• Resource graphs - per-server CPU, RAM, player count and tick-rate over time, sampled continuously, so a laggy or overloaded server jumps out at a glance
• It shows you what you're allowed to see - the whole UI adapts to each admin's role and cluster scope, so a sub-admin only ever sees the servers they actually manage
• RCON console - fire RCON commands at any server and watch the replies come back live in a web terminal
• Audit log - a filterable record of every action - who did what, and every automated event the platform fired on its own

Clusters & Servers​

• Clusters as a real thing - group servers (even across different machines) into your own clusters with a shared database, and test the DB connection right from the UI
• Add, edit, remove servers - manage each ARK server from the browser, with a launch-preview that shows you the exact command line before you ever start it
• Full lifecycle from the web - start, stop, restart, install and update any server, with graceful countdowns and live phase-by-phase progress
• Bulk cluster actions - start, stop, restart, update, back up, push configs or refresh plugins across every server in a cluster in one click
• INI editor - edit GameUserSettings / Game INI at cluster or server scope in the browser, pull the current live file off disk, and flip a server between its own INI and the cluster's shared one
• Mod-config manager - browse, upload, edit and delete mod config file trees per cluster or per server, with servers inheriting the cluster's tree by default
• Player management - pull the live player list and kick or ban straight from the UI
• Logs & save-games - list, read, tail and download server + API logs, and browse the save-game folder tree without touching the box

Plugins​

This is where the manager really earns its keep - a proper plugin pipeline instead of copying DLLs by hand.

• Central plugin library - upload and catalog your AsaApi plugins (DLL, PluginInfo, PDB and hot-reload .arkapi) in one shared place
• Version history & rollback - every upload is versioned; review past builds, leave notes, set how many to keep, and roll back to any of them
• Smart DLL push - push a plugin to every server that uses it and it works out how to deliver: a live hot-reload for online servers, a cold drop for offline ones, and a fresh install for a server that's never had it - with a pre-flight that shows you the online / offline split first
• Fast, cached deploys - plugins are staged to each machine once and then applied locally to as many servers as you like. Pushing 30 plugins to 10 servers stops being an hour of waiting and becomes one sync plus a handful of local copies
• Layered config resolver - plugin config resolves per-file down a scope chain (rotation > server > machine > group > cluster > global), with the most specific scope winning for each filename - and a preview that shows you exactly which scope won
• Config drift view - a fleet-wide look at where a server's deployed config has drifted from what it's supposed to be
• Plugin groups - define named groups of servers to share a config layer that sits between cluster and machine scope
• Per-server plugin selection - choose which plugins run on each cluster, override per server, pull one plugin off a single server and put it back later
• Auto-discovery & pull-back - scan a running server for what's actually deployed and pull any plugin straight back into the library
• Push configs only - deploy just the resolved config files (whole plugin or one group), or shove a single server's config to disk on demand

Mods​

• Mod catalog & assignment - track your mods and assign them to clusters and servers; the launch mod list builds itself
• CurseForge update tracking - the Host polls CurseForge for new mod files and drops a Discord notification - with the changelog, per affected cluster - the moment a tracked mod updates
• Wildcard event mods - when a seasonal event mod drops a new file, it auto-adds to your opted-in servers, works out the event's end date, warns 24 hours before removal and pulls itself back out when the window closes - all announced in Discord
• Clean only what changed - on start, either force ARK to re-pull every mod fresh, or surgically delete just the mod folders that are out of date so ARK re-fetches only those

Backups & Wipes​

• Backups, on demand or scheduled - timestamped per-server backups of saves, config and plugins, with live progress, and every graceful shutdown can leave a restore point automatically
• Map-aware & self-pruning - backs up only the save files for the active map, and prunes old archives by age and count so storage doesn't run away
• Browse, download, peek, restore - list a server's backups, download any archive off the box, peek inside to see which players and tribes it holds, then restore the whole thing - or surgically, down to a single player or tribe
• Wipe profiles - reusable wipe recipes: which servers to stop and wipe, which save files / SavedArks to clear (with wildcard exclusions), which cluster DB tables to drop, and which seed files to copy back in afterward
• Scheduled wipes - run a wipe profile automatically, with a preview of what gets dropped and a warning countdown before it fires, plus a full history of every wipe run

Scheduling & Automation​

• Scheduled tasks - recurring jobs (nightly restarts, updates, config pushes, cache checks, backups) at server or cluster scope, with run-now, a run history and proper host-timezone handling
• Conflict-safe - a scheduled restart quietly stands down if it would collide with an imminent map rotation on the same server, so the two never fight
• Map rotation for event servers - opt a server into rotation and it cycles through a map pool on a schedule: swaps the map, deep-merges the per-map plugin config overrides, rebuilds the launch bat, pushes configs and restarts with a countdown. Define the pool, reorder it, set per-map overrides, preview the next map or rotate now

Discord​

• Sign in with Discord - admins authenticate through Discord, gated by the guilds and owner IDs you allow
• Control panels - pinned messages the bot keeps live with real buttons for cluster- and server-level controls, refreshed on a loop and whenever something happens
• Status embeds - auto-updating cluster / server status embeds posted to your channels
• Event routing & templates - send different event types (updates, wipes, restarts, mod updates and more) to the channels you choose, with message templates you can customise and preview
• Guild & role sync - register your guilds, sync their roles, and use them for permission mapping

Users, Roles & Permissions​

• Role-based access - owner-managed roles built from a granular permission catalog, with role cloning and per-role editing
• Cluster-scoped - scope a role's permissions to specific clusters, so an admin only ever manages their own servers
• User management - approve or reject new sign-ins, assign roles, disable and remove users

SteamCMD, Machines & Maintenance​

• SteamCMD cache per machine - status, install and cache-update controls per box, plus a one-click update-all-caches across the whole fleet
• Build checks - force a fresh installed-vs-latest ARK build check per machine, with Discord notifications for up-to-date / update-available / update-started
• Auto-update on new build - a "check and maybe update" mode that updates servers automatically when a new build lands, on demand or on a schedule
• Shared asset manager - one central store for AsaApi, the ASA loader, 7-Zip and the manager plugin, fetched or uploaded once and pushed out to every server
• Machine registry - add machines by name / IP / port / secret; the Host handshakes the agent and shows live online / offline from its heartbeat, with secrets encrypted at rest
• Update Certificates - a per-machine button that installs the Amazon Trust certificates ARK needs for mod downloads (and replaces stale ones) - the fix for "mods won't download on a fresh Windows box"
• Restart Client / Host from the web - remotely bounce a machine's agent or the Host itself, with a liveness probe so the page reconnects on its own when it's back

Under the Hood - the Client Agent​

The Client is the muscle. You never open it directly - the Host drives it - but here's what it's doing on each machine so you know your servers are in safe hands.

Install & Update​

• SteamCMD install / reinstall - installs a full ASA server from scratch, wires up the launch script and loader, and can bootstrap SteamCMD itself on a bare Windows box
• Delta-copy updates - mirrors only the files that actually changed from the local cache, so an update finishes in a fraction of a full recopy
• Live asset refresh - drops a new API version or fresh plugins onto an installed server without a full stop-and-recopy - a multi-minute update becomes a few seconds
• Version awareness - reports the ARK version a server is running, the build it has, and whether that matches the cached build

Lifecycle & Safety​

• Graceful shutdown - broadcasts an in-game countdown, saves the world, backs up if enabled, kicks everyone, then exits cleanly - and short-circuits to save-and-exit if the server's already empty
• Restart & update-then-start - stop-and-relaunch as one operation, with the option to leave a server down after an update
• Multi-stage force-kill - when a server won't die, it escalates through progressively stronger kills (standard, elevated-privilege, then a direct low-level terminate), verifying between each, so even a hung save-in-progress gets cleared
• Ghost-window cleanup - hunts down the leftover console windows, taskbar ghosts and orphaned helper processes a force-kill leaves behind, so a stopped server is truly gone
• Responsive at scale - starts and stops whole clusters at once without the agent freezing, so big multi-map operations stay fast
• Live status - reports exactly which phase a stop / restart / update is in (counting down, saving, backing up, kicking, force-killing) so you watch it happen

Watchdog & Recovery​

• Crash watchdog - watches every running server and auto-relaunches it if it crashes or vanishes, telling the Host so it's logged and announced - with a startup grace period so boot never triggers a false alarm, and a per-server toggle for maintenance
• Resumes interrupted work - if the agent itself restarts mid-shutdown or mid-update, it picks the operation back up on boot instead of leaving a server half-done
• Servers survive agent restarts - running servers keep running when the agent bounces; it just re-attaches
• Crash artifacts - collects the latest server + API logs and the crash stack after a crash so the Host can attach them for diagnosis

Locked Down​

• Path allowlist - refuses to read or write outside the folders you've explicitly authorised, with a hard block on Windows, Program Files, user profiles and drive roots even if the allowlist is misconfigured
• No shell injection - every external command runs as a discrete argument array, never a shell string, and paths are scrubbed for metacharacters
• Per-machine secret - every single request is checked against a shared secret, so only the paired Host can command the agent
• Isolated launches - ARK servers start as fully detached processes that don't inherit the agent's network socket

Under the Hood - the In-Game Plugin​

A small native AsaApi plugin runs inside each ARK server. It's what makes the dashboard's status actually true - and it handles the in-game countdown banners your players see.

• Live server state - writes RCON connection info, ServerKey, map name, uptime, player count, PID and status into the cluster database on a heartbeat, so the manager always knows the real state of every server
• Real online detection - watches the server's own log for the "advertising for join" marker before it reports online, so starting means starting and online means players can actually connect
• In-game countdowns - shutdown, restart and update countdowns broadcast as on-screen banners that update in place each second (no more spamming a new message every tick), with the final message replacing the timer when it hits zero
• Styled & per-message - every message has its own text colour, background, size, duration, screen position and optional copy to chat, each configured independently
• Forced RCON IP - honours the per-server RCON IP you set in the manager instead of auto-guessing one, so the manager always talks to the right address

Plugin Commands​

| Command | Where | Description |
|---------|-------|-------------|
| ShadowManager.ShutdownCounter <minutes> <reason> | Console / RCON | Displays a shutdown countdown to players |
| ShadowManager.RestartCounter <minutes> <reason> | Console / RCON | Displays a restart countdown |
| ShadowManager.UpdateCounter <minutes> <reason> | Console / RCON | Displays an update countdown |
| ShadowManager.KickAllPlayers [minutes] [message] | Console / RCON | Warns for the given minutes, then kicks everyone at zero |
| ShadowManager.CancelCountdown [type] | Console / RCON | Cancels an active countdown (one type, or all) |

Multiple countdowns of different types can run at once; issuing the same type again replaces the previous one with a cancellation notice. All of this is normally driven for you by the Host - the commands are there for manual control too.

In Short​

Support/Questions​