Configuration Reference

Every setting that actually exists on the 46 Glacier Pack addon panes, grouped per addon with an anchor each. Settings marked (env-only) live in the addon's config file and are tuned via .env, not the hub. Addons that manage records instead of settings (rules, destinations, alerts) list those fields under Managed data.

TIP

All saves go through the hub's _hub round-trip: you always land back on the same pane, with validation errors shown above the form when a value is rejected.

Egg-name gating (shared by all game addons)

Every game-specific addon ships with a default egg-name gate: it only appears on servers whose egg name matches its shipped pattern (ARK tools on ARK eggs, Hytale on Hytale, FiveM on FiveM, FastDL on Source-engine games, Bedrock addons on Bedrock eggs, the Minecraft Java tools on Java eggs). Two settings control this on each addon's pane:

Setting	Type	Default	What it does
egg_name_gating	checkbox	on	Master switch for name-based gating. Off = the addon shows on every egg.
egg_name_pattern	text (regex)	per-addon pattern	Case-insensitive regex matched against the server's egg name (IDs differ per install, names don't). A live hint under the field shows how many of the panel's eggs match as you type.

Addons that already had an egg-ID allowlist keep it — an ID match always wins, even when the name pattern doesn't. Server Wiper is the exception to the family rule: the wiper itself is universal and only its Rust map library gates to Rust eggs.

---

Basic pack

Audit Log

No settings form — the pane is a filterable viewer (full-text search, event scope, actor type, date range) with a per-entry detail dialog and retention tools (prune entries older than N days, or clear the log; both behind confirmation prompts).

Setting	Type	Default	What it does
per_page (env-only)	number	25 (AUDIT_LOG_PER_PAGE)	Entries listed per page (clamped 1–100).

Common mistakes: treating prune/clear as reversible — both delete activity rows permanently; expecting entries the panel never recorded — the addon reads the stock activity_logs feed and adds no extra logging of its own.

---

Auto Suspend

Sub-tabs: Overview (counters + latest activity), Servers (per-server expiry dates, one inline form per row), History, Settings. Settings save to admin.auto-suspend.settings.update.

Setting	Type	Default	What it does / when to change
suspension_enabled	checkbox	on	Master switch — off pauses warnings, suspensions, lifts and deletions.
warn_days	number (0–60)	3	Days before expiry the owner receives one warning email; 0 disables it.
auto_delete_enabled	checkbox	off	Delete servers suspended by this addon once they outstay the grace period.
auto_delete_days	number (1–365)	14	Grace period in days (suspended) before auto-delete fires.
excluded_nodes[]	multi-select	empty	Nodes whose servers are never auto-deleted.
retention_days	number (7–365)	90	History rows older than this are pruned daily.

Managed data: per-server expiry dates on the Servers tab. Extending a suspended server's expiry into the future lifts the suspension on the next scheduler pass.

Common mistakes: missing the panel cron — processing runs every five minutes via schedule:run (php artisan auto-suspend:process by hand); warning emails silently absent when the panel's mail settings are broken; expecting manual suspensions to be auto-deleted — only addon-applied suspensions are eligible.

---

Command History

Sub-tabs: Overview (capture status, totals, per-server counts, the 25 most recent commands panel-wide, per-server purge and purge-everything) and Settings.

Setting	Type	Default	What it does / when to change
enabled	checkbox	on	Pause/resume capture panel-wide.
max_per_server	number (50–1000)	200	Entries kept per server; excess is pruned on save.
list_limit (env-only)	number	50 (COMMAND_HISTORY_LIST_LIMIT)	Newest entries shown in the console dropdown.

Both UI settings also have env defaults (COMMAND_HISTORY_ENABLED, COMMAND_HISTORY_MAX_PER_SERVER).

Common mistakes: expecting the dropdown to show the full retention — it always shows the newest list_limit entries; purges (per-server or global) have no undo.

---

Config Editor

No settings form — this pane edits the panel's live .env directly. Sub-tabs: Editor (structured, grouped by prefix) and Raw file (whole-file textarea).

Field	Type	Default	What it does
env[<KEY>] (per parsed key)	text / password / true-false select	current .env value	One input per env key. Secret-looking keys (PASSWORD, SECRET, KEY, TOKEN, SALT) are masked with a reveal toggle; boolean keys get dropdowns.
env[APP_KEY]	password	always blank	Write-only. Blank keeps the current key; pasting a new one rotates it (format-validated as Laravel base64: or 32-char).
contents (Raw tab)	textarea	entire file verbatim	Bulk raw edit with the same validation, backup and atomic-write flow.
backups_keep (env-only)	number	10	Timestamped .env snapshots kept in storage/app/config-editor/backups.

Common mistakes

• Editing APP_KEY. Never. Panel data encrypted with it (node database passwords) becomes unreadable.
• Changing APP_NAME, SESSION_*, CACHE_PREFIX or REDIS_* casually — it invalidates every active session panel-wide, logging everyone out immediately.
• Forgetting that backups contain full .env secrets — they are written 0640 in a 0750 directory; treat them like the .env itself.

---

Console Log Share

Single settings pane (no sub-tabs). Posts to admin.console-log-share.update; Reset to Defaults restores the shipped values.

Setting	Type	Default	What it does / when to change
paste_endpoint	url	https://api.mclo.gs/1/log	Any mclogs-compatible API: the panel POSTs the log as form field content and expects JSON containing url. Change when self-hosting a paste service.
upload_enabled	checkbox	on	Off hides the Share button and rejects uploads; Copy Log stays available.
placement	select	top (Above the console)	Where the Copy/Share bar renders on the console page.
default_paste_endpoint (env-only)	—	CONSOLE_LOG_SHARE_PASTE_ENDPOINT, same URL	Shipped default the reset button returns to.
max_log_lines (env-only)	number	1500 (CONSOLE_LOG_SHARE_MAX_LINES)	Trailing console lines requested from the node.
max_upload_bytes (env-only)	number	5242880 (CONSOLE_LOG_SHARE_MAX_BYTES)	Largest payload pushed to the paste service; older lines are trimmed beyond it.

Common mistakes: pointing paste_endpoint at a plain HTML pastebin (it must be the mclogs-compatible API URL); changing env overrides without php artisan config:clear.

---

Login Activity

No settings form. Sub-tabs: Logins (full history with user / date-range filters, pagination) and Sessions (live mirrored sessions with user filter, per-row revoke and revoke-all for one user). Deleted users show as "Deleted user #id".

Setting	Type	Default	What it does
admin_per_page (env-only)	number	25 (LOGIN_ACTIVITY_ADMIN_PER_PAGE)	Rows per page on the admin overview (clamped 1–100).
client_per_page (env-only)	number	8 (LOGIN_ACTIVITY_CLIENT_PER_PAGE)	Rows per page in the account-area widget.
sessions.enabled (env-only)	bool	true (LOGIN_ACTIVITY_SESSIONS)	Mirror and track live sessions for the Sessions tab and self-service revoke.
geo.enabled (env-only)	bool	true (LOGIN_ACTIVITY_GEO)	Coarse ipwho.is geolocation per unique IP; set false to keep IPs in-house.
geo.cache_days (env-only)	number	30 (LOGIN_ACTIVITY_GEO_CACHE_DAYS)	One lookup per unique IP per cache window.
proxy.detection (env-only)	—	heuristic (LOGIN_ACTIVITY_PROXY_DETECTION)	Network-name keyword heuristic; set proxycheck for an authoritative verdict.
LOGIN_ACTIVITY_PROXYCHECK_KEY (env-only)	—	empty	Free-tier proxycheck.io key, only used with the proxycheck detector.

Common mistakes: expecting authoritative VPN detection out of the box — the default heuristic is best-effort (set the proxycheck mode for a real verdict); a revoked session is not instant-logout down to the millisecond — it dies on its very next request. History is kept on uninstall unless you pass --drop-table to remove.sh.

---

Move Files

No settings at all — the pane is an informational card. The addon enhances every server's file manager automatically once installed; there is no config file and no backend.

Common mistakes: forgetting a hard refresh (Ctrl+Shift+R) on the Files page after install, and forgetting to re-run install.sh after panel updates overwrite templates/base/core.blade.php.

---

Node Status

Sub-tabs: Overview (settings + per-node overrides), Monitors, Sharing & Alerts, Service Updates. Settings save to admin.uptime.settings.

Setting	Type	Default	What it does / when to change
enabled	checkbox	on	Master switch — off hides the Uptime tab and stops collection.
poll_interval	select	60 s (30 / 60 / 120 / 300)	How often the panel polls Wings. Lower = fresher graphs, more node load. The collector honors the configured interval.
retention_days	select	90 (30–365)	Nightly pruning horizon; the graph window maxes at 365 days.
users_see_node_uptime	checkbox	on	Whether regular users see node health data.
show_node_names	checkbox	off	Off masks real node names/FQDNs for users ("Your Node", "Node 1"); admins always see real names.
public_page_enabled	checkbox	off	Unauthenticated /status page plus JSON feed (/api/uptime-public/summary), rate-limited and noindex.
public_page_title	text	shipped string	Title of the public status page.
public_show_uptime / public_show_latency / public_show_monitors / public_show_notices	checkboxes	on	Per-element display toggles for the public page.
public_show_node_names	checkbox	off	Show real node names publicly — node FQDNs are never exposed either way.
discord_webhook_url	url	empty	Discord webhook for down/resolved incident embeds.
discord_alert_nodes / discord_alert_monitors	checkboxes	on	Alert categories: node and monitor transitions.
discord_alert_servers	checkbox	off	Also alert on per-server transitions (noisy on big panels).
discord_hide_endpoint	checkbox	off	Mask the endpoint in alerts posted to public channels.
job_timeout / node_connect_timeout / monitor_timeout / cache_ttl (env-only)	numbers	45 / 3 / 6 / 300	Collection job timeout, per-node and per-monitor connect timeouts, cache TTL (seconds).

Managed data: per-node tracking overrides (toggle per node); custom Monitors (name, type — http / tcp / udp, target, port), each with live state, 24 h uptime, latency, last check, failure reason and pause/resume; Service Update posts with title (required), event_at (optional — renders a live countdown) and body_html (required; whitelisted basic HTML). A Send test alert button verifies the Discord webhook.

Common mistakes: missing the panel cron — collection (uptime:collect) and pruning run via the scheduler; forgetting the uptime.read subuser permission, without which users see no Uptime tab; reading UDP monitors as proof of life — a silent host counts as up unless the network refuses the packet (noted in the UI).

---

PWA

Single settings pane (posts to admin.pwa.update) with live endpoint links; Reset to Defaults restores the shipped values, disables the shell and rolls client caches on the next page load.

Setting	Type	Default	What it does / when to change
enabled	checkbox	off	Master switch. Disabling makes every client unregister the service worker and purge its caches on the next page load.
app_name	text (max 80)	empty (panel name)	Shown in the install dialog and on the home screen.
app_short_name	text (max 30)	empty	Used where space is tight (under the home-screen icon).
app_description	textarea (max 200)	empty	Shown in some install dialogs.
theme_color / background_color	color	#0e4688	Browser toolbar color / splash-screen background.
display	select	standalone	Display mode; standalone is the classic app-window mode.
orientation	select	any	Orientation lock.
status_bar_style	select	default	iOS status bar style (Apple devices only).
icon_192 / icon_512 / icon_maskable	text	panel favicons	Paths (/favicons/…) or full URLs; empty slots fall back to the bundled favicons.
cache_strategy	select	cache-first	cache-first (fastest) or network-first (freshest) for static assets.
cache_max_age	number (1–720)	24 h	Cached entries older than this refresh from the network.
cache_assets	checkbox	on	Cache panel JS/CSS/fonts/images.
cache_api	checkbox	off	Network-first caching of /api/client GET responses with a cached fallback.
offline_enabled	checkbox	on	Branded offline fallback page at /pwa/offline.
offline_title / offline_message	text	shipped strings	Offline page copy.
prompt_enabled	checkbox	on	Dismissible install banner; dismissed banners stay hidden seven days, iOS shows Share → Add to Home Screen instructions.
prompt_delay	number (0–600)	30 s	Delay before the banner appears.
prompt_button / prompt_title / prompt_text	text	shipped strings	Banner copy.

Common mistakes: testing over plain HTTP — service workers and install prompts require a secure context (localhost is exempt); enabling cache_api and being surprised by briefly stale data on flaky connections.

---

Panel Logs

No settings form — a tool pane with log files grouped by source and a tail viewer (level filters, 2/5/10/30 s auto-refresh, download, clear, delete). The panel's storage/logs is joined by auto-detected Nginx (/var/log/nginx) and Wings (/var/log/pterodactyl) sources plus any you configure.

Setting	Type	Default	What it does
PANEL_LOGS_TAIL_LINES (env-only)	number	400	Lines of context loaded when a log opens.
PANEL_LOGS_SCAN_BYTES (env-only)	number	4194304 (4 MiB)	Backwards scan budget for the initial tail.
PANEL_LOGS_POLL_MAX_BYTES (env-only)	number	524288 (512 KiB)	Max bytes per poll response.
PANEL_LOGS_POLL_INTERVAL (env-only)	number	5 (seconds)	Default auto-refresh interval in the viewer.

Log sources (config/panellogs.php → sources): each source takes label (display name), path (absolute directory), pattern (glob, default *.log), parser (laravel, auto = Laravel + Nginx error + Wings, or plain), readonly (default true — hides and refuses clear/delete server-side) and auto (default false — silently skip when the directory is missing). Run php artisan config:clear after editing.

Common mistakes: expecting clear/delete on external sources — they are read-only by design so the addon can never truncate web server or daemon logs; a source whose directory exists but is unreadable by the panel user shows a permission notice — grant read access (e.g. group membership) to list it.

---

Quick Files

Single pane: settings, usage stats (total pins, servers with pins, most pinned servers) plus two separate actions — Reset to defaults and Clear every pin panel-wide.

Setting	Type	Default	What it does / when to change
enabled	checkbox	on	Off hides the Quick Access bar and row stars panel-wide.
max_pins	number (0–500)	25	Pins per server; 0 = unlimited.

Common mistakes: forgetting pins are server-wide — everyone with file access sees the same bar; worrying about the re-check button — it only drops pins whose target is confirmed missing, unlistable directories are left untouched.

---

Recycle Bin

Single pane: global settings, storage overview, top-servers table and per-egg overrides.

Setting	Type	Default	What it does / when to change
enabled	checkbox	on	Off makes deletes permanent again (stock behavior).
retention_hours	number (1–8760)	24	Hours a trashed file stays recoverable before purge.
default_max_mb_per_server	number (0–10485760)	0 (unlimited)	Per-server trash capacity in MB, counted against the server's real disk. Set this on shared nodes. Files too large for the bin can only be deleted permanently.
max_files_per_delete	number (1–10000)	100	Cap per delete request.
purge_batch_size	number (50–5000)	200	Rows per chunk processed by the hourly purge job.
logging_enabled	checkbox	on	Write activity-log entries for trash / restore / purge.
protect_trash_dir	checkbox	on	Protects .trash from manual moves and deletion.
reroute_permanent_delete	checkbox	off	Route "delete permanently" requests (UI or API clients sending the permanent flag) through the bin too; purges from inside the bin still destroy files.

Managed data: per-egg overrides — egg_id (required), retention_hours (blank inherits global), max_mb (blank inherits global).

Common mistakes: leaving capacity unlimited on busy nodes (trash eats real disk); not re-running install.sh after panel updates overwrite the patched core files. The client UI integrates at runtime, so no yarn build is needed (a compiled SPA integration remains optional).

---

Resource Alerts

Sub-tabs: Overview, Rules, History, Settings. Settings save to admin.resource-alerts.settings.update; Send Test Alert emails all admins and hits the configured webhooks.

Setting	Type	Default	What it does / when to change
monitoring_enabled	checkbox	on	Pause switch — off skips every poll and fires nothing.
discord_webhook_url	url	empty	Discord incoming webhook for alert delivery.
slack_webhook_url	url	empty	Slack incoming webhook.
retention_days	number (1–365)	30	History records older than this are pruned daily.
history_per_page (env-only)	number	25	History pagination size.

Managed data (rules): name (max 64), metric (cpu, memory, disk, network_rx, network_tx), operator (above/below), threshold (default 90; % for cpu/memory/disk, Mbps for network), scope (global/node/server), node_id/server_id (per scope), duration_minutes (1–60, default 1 — consecutive 1-minute samples outside the threshold before firing), enabled, notify_mail, notify_webhook. Layering: global rule → per-node override → per-server override; most specific wins.

Common mistakes: enabling notify_webhook on rules before saving webhook URLs; treating CPU/memory/disk thresholds as absolute values (they are percentages of the server's limits); missing the panel cron — sampling runs once per minute via the scheduler.

---

Schedule Presets

Preset CRUD plus a global export switch on one pane.

Setting	Type	Default	What it does / when to change
export_enabled	checkbox	on	Allow user-side JSON export/import of schedules.
max_preset_tasks (env-only)	number	10 (SCHEDULE_PRESETS_MAX_TASKS)	Task cap per preset.
max_import_bytes (env-only)	number	16384 (SCHEDULE_PRESETS_MAX_IMPORT_BYTES)	Largest accepted JSON import document.

Managed data (presets): name, five cron fields (with a quick-pick for common timings, validated server-side), an offer to users toggle, an only run when online flag, and an ordered task list — each task is a stock task type (send command, send power action, create backup with ignored-files payload) with a time offset (0–900 s) and a continue-on-failure flag.

Common mistakes: backup tasks are refused when the server's backup limit is 0, and applies honor the panel's per-schedule task limit — both surface as clear errors to the user.

---

Server Timezone

Single pane: master switch, label, default zone, and every server override with a per-row reset.

Setting	Type	Default	What it does / when to change
enabled	checkbox	on	Off hides the client tab panel-wide.
nav_label	text	Timezone	Label of the tab in the server navigation.
default_timezone	select	UTC	Zone used by every server without an override.

All three also have env defaults (SERVERTIMECHANGER_ENABLED, SERVERTIMECHANGER_NAV_LABEL, SERVERTIMECHANGER_DEFAULT_TIMEZONE).

Common mistakes: expecting a running container to pick up the zone — Docker environment changes only at the next container (re)creation, so restart after changing; container images without tzdata fall back to UTC-style behavior (every standard yolk ships it).

---

Server Wiper

Sub-tabs: Overview, Executions, Settings.

Setting	Type	Default	What it does / when to change
nav_label	text	Wiper	Client tab label.
default_timezone	select	UTC	Preselected zone for new schedules.
max_schedules_per_server	number	10	Schedule cap per server.
allow_live_wipes	checkbox	on	Permit wipes without stopping the server first.
retention_days	number	30	Execution history kept.
stop_grace_seconds	number	15	How long to wait for the server to go offline before wiping.
rust_eggs / ark_eggs	text	empty	Comma-separated egg IDs that unlock the Rust/ARK options.
rust_map_sizes	text	1000,1500,…,6000	Selectable map sizes.
rust_default_map_size	number	4500	Preselected map size.

Managed data: per-server wipe schedules and map libraries (up to 50 .map URLs) live on the client side; the Executions tab lists every run with status, files deleted, power actions, commands sent and applied Rust variables.

Common mistakes: a glob that matches a folder removes the folder whole — test patterns on a throwaway server first; each schedule runs in its own timezone, not the panel's.

---

Startup Presets

Sub-tabs: Presets (CRUD with scope badges) and Settings.

Setting	Type	Default	What it does / when to change
custom_command_enabled	checkbox	off	Let users with startup.update type a fully custom startup command instead of only picking presets.
custom_command_nest_ids[] / custom_command_egg_ids[]	multi-selects	empty (every server)	Restrict free-form editing to specific nests and/or eggs.

Managed data (presets): name, optional description, the startup command (egg placeholders like are substituted at server start), and nest/egg scoping (empty = offered everywhere).

Common mistakes: none major — users can always reset to the egg's original startup command from the same picker.

---

URL Download

Single settings pane (one form, four boxes). Reset to defaults and Clear caches (remembered URLs + activity) are separate actions.

Setting	Type	Default	What it does / when to change
max_file_size_value + max_file_size_unit	number + select (MB/GB)	10 GB	Files larger than this are rejected before download. 0 = unlimited.
allowed_url_schemes	text	http,https	Comma-separated scheme allow-list.
allowed_file_extensions	text	empty (all allowed)	Comma-separated without dots, e.g. zip,gz,tar,jar.
max_urls_allowed	number (1–100)	3	URLs per batch.
max_concurrent_downloads	number (1–50)	3	Queued + active downloads per user per server.
max_query / max_query_window	numbers	10 / 60	URL-validation rate limit (queries per window, seconds).
button_label	text (max 60)	Download from URL	File-manager toolbar button label.
show_progress	select	popup	popup, deck (modal + floating deck), deck_only, none. The deck survives page reloads.
memory_limit	number (1–20)	20	Remembered URL chips per user.
allow_simultaneous_downloads	checkbox	on	Let users run a batch simultaneously instead of queued.
prevent_incomplete_downloads	checkbox	on	Post-transfer verification; empty/missing results are flagged and cleaned up.
follow_redirects	checkbox	on	Follow HTTP redirects when probing/downloading.
allow_unknown_size	checkbox	off	Accept downloads whose size the HEAD probe cannot determine — size limits cannot be enforced for these.
validate_urls_on_entry	checkbox	on	Validate URLs as the user types.
prevent_html_downloads	checkbox	off	Reject URLs that resolve to HTML pages.
max_active_history / max_failed_history / max_completed_history	numbers (0–1000)	20 each	History rows kept per state; 0 = unlimited.
stale_download_timeout	number (60–86400)	3600 s	A download with no progress this long is marked failed.
probe_timeout (env-only)	number	15 (URL_DOWNLOAD_PROBE_TIMEOUT)	Seconds per remote HEAD probe.

Common mistakes: enabling allow_unknown_size and expecting size limits to still apply; forgetting transfers run on the node via the native files/pull endpoint — a broken panel↔Wings link fails every download.

---

Advanced pack

Ark Mod Installer

Single pane: the CurseForge API key, catalogue cache counters and a flush action.

Setting	Type	Default	What it does / when to change
curseforge_api_key	text	empty	Unlocks the Survival Ascended catalogue browser; Survival Evolved works keyless.
max_mods (env-only)	number	100 (ARK_MOD_INSTALLER_MAX_MODS)	Mod id cap per server.
ini_path (env-only)	path	ShooterGame/Saved/Config/LinuxServer/GameUserSettings.ini	Where ActiveMods= is written for eggs without a mod variable.
http_timeout / cache TTLs (env-only)	seconds	15; search 600, details 86400, files 1800, icons 86400	ARK_MOD_INSTALLER_HTTP_TIMEOUT, …_SEARCH_TTL, …_DETAILS_TTL, …_FILES_TTL, …_ICON_TTL.

Common mistakes: expecting mod files to pass through the panel — installing only edits the startup variable (or GameUserSettings.ini) and the game server downloads the mods itself on the next boot; ini-mode ASE servers need the node reachable from the panel (Wings file API).

---

Arma Reforger Tools

Single pane: settings, cache statistics and a one-click cache flush.

Setting	Type	Default	What it does / when to change
nav_label	text	Reforger	Client tab label.
config_file	text	config.json	Default server config file the editor opens.
addons_path	text	addons	Path the admin-tools integration resolves addons under.
admin_tools_base_path	text	profile/profile/Misfits_Logging	Base path of the in-game admin mod's JSON documents.
workshop_site	url	https://reforger.armaplatform.com	Public workshop site used for catalogue browsing.
workshop_build_id	text	auto-detected	Override for the catalogue's rotating build id (auto-detect caches 6 h and re-detects on failure).
http_timeout	number	15 s	Catalogue request timeout.
cache_ttl_minutes	number	360	Workshop catalogue cache lifetime (panel database).
config_editor_enabled / admin_tools_enabled / collections_enabled	checkboxes	on	Per-feature toggles.
collection_limit_per_user	number (1–100)	5	Saved mod-list collections per user.
webhook_username / webhook_avatar_url	text	Server Panel / empty	Identity for per-server Discord webhooks.
max_config_bytes (env-only)	number	524288 (ARMA_REFORGER_MAX_CONFIG_BYTES)	Largest config file the editor touches.

Common mistakes: a save refused after the file changed on disk is the optimistic-concurrency guard — reload and re-apply instead of expecting a silent overwrite; the panel never downloads mod binaries — the dedicated server fetches everything listed in game.mods at start (upstream behavior).

---

Backup Pro

Sub-tabs: Overview, Destinations, Backup Rules, Backups, Archives, Activity, Settings.

Settings tab (admin.s3backuppro.settings.update):

Setting	Type	Default	What it does / when to change
default_retention_count	number (0–10000)	0	Copies kept for manual offloads not attached to a rule; 0 = unlimited.
show_user_ui	checkbox	on	"Protected by S3" banner and cloud badges on the user's backup page.
download_source	select	auto	Default download/restore source: auto (local first, S3 fallback), local, cloud.
allow_user_cloud_download	checkbox	on	Users may download backups from S3 copies.
allow_user_cloud_restore	checkbox	off	Users may restore from S3 copies — overwrites server files, hence off.
auto_retention_mode	checkbox	off	Match S3 retention to each server's panel backup limit; ignores per-rule "Keep newest N".
backup_rate_limit_count / backup_rate_limit_seconds	numbers	2 / 600	Backup creation throttle; window 0 = no throttle.
notify_webhook_url	text	empty	Discord-compatible webhook for notifications; blank disables.
notify_level	select	failures	none, failures (recommended), all.
sync_check_interval	number (1–1440)	60	Minutes between sync checks.
allow_user_sync_trigger	checkbox	off	Makes the red "pending" cloud icon clickable for users.

Managed data:

• Destinations — name, access_key_id, secret_access_key (encrypted at rest; blank on edit keeps it), provider_hint (custom/aws/r2/b2/wasabi/minio/datalix — auto-fills defaults, does not affect functionality), bucket, region (default us-east-1; R2 = auto), endpoint (blank = AWS), prefix, use_path_style (on; most non-AWS providers require it), is_enabled. One-click connection tests, per-row or "test all".
• Backup rules — name, target_type (all/node/servers), node_id or servers[], destinations[], backup_mode (mirror / schedule / both / off), interval_minutes (default 1440), retention_count (default 5, 0 = unlimited), retention_days (0 = no age limit), offload (delete node copy after upload), lock_backups, ignored_files (one glob per line, .pteroignore syntax).

Env-only tuning (config/s3backuppro.php): S3BACKUPPRO_QUEUE_CONNECTION / S3BACKUPPRO_QUEUE_NAME (default standard), multipart threshold/part size (16 MiB), completion_timeout_seconds (3600), stale_after_minutes (180), max_servers_per_run (100), prune_time (04:00), log_retention_days (60).

Common mistakes: uploads sitting pending with a red cloud icon — almost always a missing queue worker on the standard queue; wrong path-style/endpoint combo for the provider (AWS: blank endpoint, path-style off; R2/B2/Wasabi/MinIO: provider endpoint, path-style on).

---

Bedrock Addon Installer

Single pane: settings plus the panel-wide install records with a clear action.

Setting	Type	Default	What it does / when to change
enabled	checkbox	on	Off hides the client tab panel-wide.
curseforge_api_key	text	empty	Required — the browser stays hidden from users until a key is saved; the key never leaves the panel.
allowed_types[]	checkboxes	all four	Catalogue sections users may browse: addons, resource packs, scripts, world templates.
default_type	select	addons	Preselected section.
egg_ids	text	empty (every egg)	Restrict the browser to specific egg IDs.
cache_ttl	number	300 s	Catalogue cache lifetime.

Common mistakes: behavior/resource packs activate in the world named by level-name (default world) — point it elsewhere first if you run multiple worlds; a restart is usually required before new packs take effect; record cleanup never touches pack files on servers.

---

Bedrock Version Changer

Single pane: settings plus the recorded change history with a clear action.

Setting	Type	Default	What it does / when to change
enabled	checkbox	on	Off hides the client tab and rejects installs.
allowed_types[]	checkboxes	both	Flavors users may install: Vanilla (official BDS) and PocketMine-MP.
allow_wipe	checkbox	on	Permit the wipe-before-install option.
allow_zip_wipe	checkbox	on	Permit archive-and-wipe (backup-before-bedrock-<timestamp>.zip left in the server root).
egg_ids	text	empty (every egg)	Restrict the picker to specific egg IDs.
cache_ttl	number (300–86400)	21600	Version catalogue cache lifetime.
history_limit	number (0–500)	50	Change records kept per server.
http_timeout / http_retries (env-only)	numbers	10 / 2 (BEDROCK_VERSION_CHANGER_HTTP_TIMEOUT, …_HTTP_RETRIES)	Catalogue request tuning.

Common mistakes: preview/beta builds are intentionally not offered — only stable releases; nodes must reach www.minecraft.net and github.com for the actual downloads.

---

Database Manager

Single settings pane plus a recent-operations log.

Setting	Type	Default	What it does / when to change
max_export_mb	number	256	Databases estimated larger are refused before a dump starts.
max_import_mb	number	64	Upload cap for imports — the pane shows your live PHP upload_max_filesize/post_max_size; raise those first for large imports.
allow_export	checkbox	on	Users with databasemanager.export may download dumps.
allow_import	checkbox	on	Users with databasemanager.import may restore uploads.
allow_wipe	checkbox	on	Permit emptying the target schema before import.
log_activity	checkbox	on	Mirror each operation into the server's activity log.
mysqldump_binary / mysql_binary / work_directory (env-only)	—	mysqldump / mysql / storage/app/database-manager	Override via DATABASE_MANAGER_MYSQLDUMP, DATABASE_MANAGER_MYSQL, DATABASE_MANAGER_WORKDIR.
export_timeout / import_timeout (env-only)	seconds	280 / 580	Process limits.

Common mistakes: setting max_import_mb above the PHP upload limits; missing default-mysql-client on the panel host (both binaries must be reachable by the web user).

---

FastDL Manager

Single pane: navigation label, nest allow-list, the global default URL and per-node URL overrides.

Setting	Type	Default	What it does / when to change
nav_label	text	FastDL	Client tab label.
allowed_nests	text	empty (every nest)	Comma-separated nest IDs the tab is available for.
default_url	url	empty	Global download base URL covering every node without its own override.
max_config_bytes (env-only)	number	1048576	Largest server.cfg the sync reads.

Managed data: per-node URL overrides (node_id + url). Built-in game profiles: Garry's Mod, CS:GO and CS 1.6 — the profile picks the URL subdirectory and the config path; the likely profile is preselected from the server's egg.

Deliberately dropped: automatic web-server provisioning (nginx vhosts, TLS, file mirroring) — the competitor's version needs a node-side agent; this addon manages URLs and configs only, and its README ships a minimal vhost example.

Common mistakes: the rule is base URL + game directory must reach the folder holding the server's maps/, models/, sound/ and materials/ folders — with a vhost rooted at the volumes directory the base URL ends in the server <uuid>; changes apply after a restart (Source games read server.cfg at boot).

---

FiveM Utils

Single pane: tab label, egg keywords, a review of which managed variables each matching egg defines, and a one-click variable re-sync for newly imported eggs.

Setting	Type	Default	What it does / when to change
nav_label	text	FiveM Utils	Client tab label (FIVEM_UTILS_NAV_LABEL).
egg_keywords	text	fivem,txadmin,cfx	Egg-name keywords deciding where the page appears (FIVEM_UTILS_EGG_KEYWORDS).
artifact_cache_minutes (env-only)	number	360 (FIVEM_UTILS_ARTIFACT_CACHE_MINUTES)	Artifact feed cache lifetime (a static fallback ships).

Managed data: none — cards write the egg variables GAME_BUILD, TXADMIN_ENABLE / TXADMIN_PORT, ARTIFACT_URL and MYSQL_CONNECTION_STRING, which the installer adds to every matching egg automatically.

Common mistakes: ARTIFACT_URL is consumed by the egg's install script on the next (re)install — a custom egg may need a one-line reference; the addon sets the txAdmin port variable but does not open firewall ports on the node.

---

Git Source Control

Single pane: host diagnostics (git version, volume root presence/writability, API base, linked-account count), settings, and the linked-account list with one-click revoke.

Setting	Type	Default	What it does / when to change
nav_label	text (max 30)	Source Control	Client tab label.
allowed_nests	text	empty (every nest)	Comma-separated nest IDs the feature is available for.
allowed_hosts	text	github.com	Comma-separated HTTPS git hosts; tokens are only ever sent to these hosts.
allow_push	checkbox	on	Off panel-wide makes every mirror read-only.
allow_manual_url	checkbox	on	Permit cloning arbitrary HTTPS URLs on allowed hosts, not just the linked account's repository list.
volume_root / git_binary / command_timeout / api_base (env-only)	—	/var/lib/pterodactyl/volumes / git / 120 s / https://api.github.com	GSC_VOLUME_ROOT, GSC_GIT_BINARY, GSC_COMMAND_TIMEOUT, GSC_API_BASE (point the last at a GitHub Enterprise Server, e.g. https://git.example.com/api/v3).

Managed data: linked connections (provider profile ↔ panel user); revoking forgets the stored token immediately.

Common mistakes: servers whose volume is not mounted on the panel host (remote nodes without shared storage) show a clear "files not reachable" state — operations run panel-side against the volume root; pulls are fast-forward-only by design, so a diverged branch reports instead of merging.

---

Hytale Mod Installer

Single pane: the CurseForge API key.

Setting	Type	Default	What it does / when to change
curseforge_api_key	text	empty	Required — CurseForge is currently the only public repository with a Hytale section.
install_directory (env-only)	path	/mods (HYTALE_MOD_INSTALLER_DIRECTORY)	Server-relative install target.
http_timeout / cache_ttl (env-only)	seconds	15 / 900 (HYTALE_MOD_INSTALLER_HTTP_TIMEOUT, …_CACHE_TTL)	Catalogue request timeout and cache lifetime.

Common mistakes: non-Hytale eggs see an explanatory notice instead of the catalogue (the installed-mod list stays available); nodes must reach the CurseForge CDN for downloads.

---

Hytale World Manager

Single pane: the CurseForge API key plus a panel-wide managed-world count.

Setting	Type	Default	What it does / when to change
curseforge_api_key	text	empty	Required — the same key as the Hytale Mod Installer works.
worlds_directory (env-only)	path	/universe/worlds (HYTALE_WORLD_MANAGER_DIRECTORY)	Where installed worlds land.
staging_directory (env-only)	path	/.hytale-world-manager (HYTALE_WORLD_MANAGER_STAGING)	Hidden staging folder for unpacks.
http_timeout / cache_ttl (env-only)	seconds	15 / 900 (HYTALE_WORLD_MANAGER_HTTP_TIMEOUT, …_CACHE_TTL)	Catalogue request timeout and cache lifetime.

Common mistakes: activate worlds while the server is offline — Hytale rewrites its config.json while running, so activation always applies on the next start; the world the server currently loads cannot be removed.

---

Mod Installer

Single settings pane with provider status badges.

Setting	Type	Default	What it does / when to change
curseforge_api_key	text	empty	From the CurseForge for Studios console. Modrinth needs no key; CurseForge stays hidden from users until a key is saved.
default_provider	select	modrinth	Preselected repository when a user opens the Mods page.
default_loader	select	fabric	Preselected loader filter (fabric/forge/quilt/neoforge/any); users can change it anytime.
install_directory (env-only)	path	/mods (MOD_INSTALLER_DIRECTORY)	Server-relative install target.
http_timeout / version_manifest_ttl (env-only)	seconds	15 / 86400	Provider request timeout; Minecraft version manifest cache TTL.
search_cache_ttl (env-only)	seconds	300 (MOD_INSTALLER_SEARCH_CACHE_TTL)	Panel-side catalogue search/version-list cache.
job_timeout (env-only)	seconds	600 (MOD_INSTALLER_JOB_TIMEOUT)	Background download job limit.
analysis_limit (env-only)	bytes	26214400 (25 MB)	Largest jar read back for update detection.

Common mistakes: assuming CurseForge works without a key; blocking outbound HTTPS to api.modrinth.com, api.curseforge.com or launchermeta.mojang.com (panel) and the mod CDNs (nodes); CurseForge-installed jars reporting "untracked" in update checks — fingerprint matching is Modrinth-only.

---

Modpack Installer

Settings pane with health badges (zip extension, staging writable, per-platform status), usage counters and a recent-installs table, plus an Installs sub-tab (tasks running right now with phase/progress, and every pack currently tracked as installed panel-wide with its backup archive name).

Setting	Type	Default	What it does / when to change
enable_modrinth	checkbox	on	Modrinth works out of the box, no key required.
enable_curseforge	checkbox	off	Requires an API key from the CurseForge console.
curseforge_api_key	text	always rendered blank	Stored key is never shown; blank keeps it.
clear_curseforge_key	checkbox	off	Check + save to remove the stored key.
button_label	text (max 40)	Modpacks	File-manager button label.
results_per_page	number (5–50)	20	Search results per page.
cache_ttl_minutes	number (5–1440)	30	Platform response cache — respects catalogue rate limits.
max_archive_size_mb	number (10–2048)	512	Pack archives/bundles larger than this are rejected.
max_files_per_pack	number (10–20000)	2500	File-count cap per pack.
max_staged_file_mb / max_staged_total_mb	numbers	5 / 64	Caps for configs and other files shipped inside the pack archive.
plan_rate_limit / plan_rate_window	numbers	5 / 300	How often one user may prepare an install per server.
allow_clean_install	checkbox	on	Users may wipe the target directory before pack files land.
allow_zip_backup	checkbox	on	Users may pick back up, wipe, then install — current files are zipped into a modpack-backup-<date> archive left in the target directory first.

Separate actions: Clear platform cache, Clear staging files, Reset defaults (resets limits, keeps the API key).

Common mistakes: missing the PHP zip extension or a non-writable storage/app/modpack-installer (both shown as pane badges); expecting egg/Docker changes — the addon installs files only, and client-only mods plus third-party-opt-out files are auto-skipped and listed on the confirmation screen; uninstalling a pack installed before 1.1.0 — there is no file manifest, so the button explains why it cannot. A background install keeps its state in the panel cache for the plan lifetime (default 45 minutes); a cache flush mid-install settles the task as a clearly-labelled failure that can simply be started again.

---

Node Stats

Sub-tabs: Overview, Nodes, Capacity, Historical, Top Consumers, Reports, Settings.

Setting	Type	Default	What it does / when to change
collection_enabled	yes/no radio	yes	Master switch for metric collection from Wings.
collection_interval	number (10–300)	30 s	How often metrics are collected.
retention_raw_days	number (1–365)	7	Raw snapshot retention.
retention_1m_days	number (1–365)	30	1-minute aggregate retention.
retention_5m_days	number (1–730)	90	5-minute aggregate retention.
retention_1h_days	number (1–1825)	365	Hourly aggregate retention.
retention_1d_days	number (1–3650)	1825	Daily aggregate retention.
retention.aggregates_1mo_days (env-only)	number	0 (keep forever)	Monthly aggregate retention.

Env-only (config/nodeanalytics.php): queue name nodeanalytics, batch sizes (collection 50 / aggregation 500), cache TTL 60 s, reports.max_rows 100000.

Common mistakes: dashboards staying empty — the addon needs both the panel cron (the scheduler runs nodeanalytics:schedule every minute) and a Supervisor worker on the nodeanalytics queue (data/install-queue-worker.sh sets it up).

---

Permission Manager

Sub-tabs: Overview, Roles, Members, Audit Log. No global settings form — the panes manage roles, members and scoped assignments.

Managed data:

• Roles — name (required), description, pages[] (which staff-area pages the role can open: Users, Servers, Nodes, Nests), see_all_servers, see_all_users (off = only explicitly assigned servers/users), allow_build_edit (edit CPU/memory/disk/swap/IO/feature limits via the panel's own build service), subuser_permissions[] (defaults to console, power actions and file read — provisioned as a real subuser on every in-scope server; editing re-syncs all members).
• Members — user_id, role_id, active (unchecking revokes access and provisioned subusers but keeps assignments). Per member: assign specific users (staff then see those users and their servers) and specific servers.
• Config-only (config/permissionmanager.php): the page list and default_subuser_permissions (websocket.connect, control.console, control.start, control.stop, control.restart, file.read, file.read-content).

Common mistakes: trying to delete a role that still has members (refused — reassign or remove them first); uninstalling without removing members or running php artisan permgr:sync, which leaves provisioned subusers behind.

---

Player List

Single settings pane with a built-in host:port test tool (reports whether the full UDP query answered).

Setting	Type	Default	What it does / when to change
enabled	checkbox	on	Off hides the widget panel-wide without uninstalling.
protocol	select	auto	auto (query, then ping), query (UDP query only), ping (status ping only). Query returns the full list but needs enable-query=true; ping needs no setup but returns only a name sample.
avatars	checkbox	on	Head images next to player names.
avatar_url (env-only)	url template	https://mc-heads.net/avatar/{name}/40 (PLAYER_LIST_AVATAR_URL)	Avatar provider; {name} placeholder. A local placeholder head is used when the image cannot be loaded.
actions	checkbox	on	Moderation menu on player rows (kick, ban, op/deop, gamemode, whitelist add/remove) for users with console access.
timeout_ms	number (250–10000)	2000	Per probe attempt.
cache_seconds	number (0–300)	15	Per-server result cache — prevents query floods.
poll_seconds	number (5–300)	20	Browser refresh interval of the console widget.

All UI settings also have env defaults (PLAYER_LIST_ENABLED, PLAYER_LIST_PROTOCOL, PLAYER_LIST_AVATARS, PLAYER_LIST_ACTIONS, PLAYER_LIST_TIMEOUT_MS, PLAYER_LIST_CACHE_SECONDS, PLAYER_LIST_POLL_SECONDS).

Common mistakes: expecting full player lists from servers without enable-query=true and the query port reachable from the panel — the ping fallback only ever returns a sample (the card's Enable query button writes enable-query=true and query.port for users with the file-edit permission; a restart is required); avatar images load client-side, so users' browsers need outbound access to the avatar provider.

---

Player Manager

Single settings pane plus a stored-credentials table (passwords shown only as stored/—, never the value) with revocation.

Setting	Type	Default	What it does / when to change
enabled	checkbox	on	Off answers the client page and every API action with a disabled notice.
avatars	checkbox	on	Head renders served through a caching panel-side proxy (browsers never contact third-party services).
playerdata	checkbox	on	Offline roster and the player-data viewer/editor (playerdata/*.dat via the node file API).
default_port	number (1–65535)	25575	Used when neither a stored override nor a server variable defines a port (vanilla default).
connect_timeout	number (1.0–15.0)	3.0 s	RCON connect timeout.
read_timeout	number (1.0–30.0)	4.0 s	RCON response timeout.

Config-only: the server-variable names consulted when no per-server override exists — RCON_PASSWORD, RCON_PORT, RCON_HOST (overridable via PLAYER_MANAGER_VARIABLE_PASSWORD/PORT/HOST).

Common mistakes: servers unreachable because RCON is not enabled in server.properties — use the built-in Fix RCON button, which writes enable-rcon=true, rcon.port and rcon.password through the node's file API (generating and storing a password when the server has none); forgetting the credential resolution order — stored per-server override first, then server variables, host defaulting to the primary allocation; player-data edits/wipes are refused while the player is online.

---

Plugin Installer

Single settings pane (save only; no reset/test actions).

Setting	Type	Default	What it does / when to change
provider_modrinth / provider_hangar / provider_spiget	checkboxes	on	Free catalogues, no API key required. Spiget lists premium resources but cannot install them.
provider_curseforge	checkbox	off	Stays unavailable until a valid API key is saved, even when toggled on.
provider_polymart	checkbox	off	Marketplace catalogue; only entries flagged downloadable are installable.
curseforge_api_key	text	empty	From the CurseForge console; empty keeps CurseForge disabled.
install_directory	text	/plugins	Server-relative install target — /plugins for Bukkit-based, /mods for modded setups. Normalized to a leading slash; falls back to /plugins when empty, /, containing .., or over 64 chars.
analysis_max_mb	number (1–100)	15	Jars larger than this are listed in the Installed tab but skipped for metadata extraction and update checks.
provider_timeout (env-only)	seconds	20 (PLUGIN_INSTALLER_PROVIDER_TIMEOUT)	Catalogue request timeout.
catalogue_cache_ttl (env-only)	minutes	5 (PLUGIN_INSTALLER_CACHE_TTL)	Panel-side cache for search, version and details responses.

Common mistakes: no installs landing — nodes must reach the provider CDNs (Wings performs the download via native file-pull); strict CSP blocking provider icon hosts (icons fall back to letter placeholders).

---

Reverse Proxy

Sub-tabs: Settings, Whitelist (per-server limit overrides with a server search picker), All Proxies (every rule on the panel with server, owner, node, SSL and DNS status, plus delete for moderation).

Setting	Type	Default	What it does / when to change
allowed_nodes[]	checkboxes	empty (all nodes)	Restrict the feature to specific nodes.
default_proxy_limit	number (0–1000)	3	Proxy rules each server may create.
whitelist_enabled	checkbox	off	Whitelist mode gives individual servers their own limit.

Managed data: whitelist entries (server, proxy_limit 0–1000). Client rules carry domain (unique panel-wide), allocation_id, ssl_type (none / letsencrypt / custom) plus ssl_email or ssl_certificate + ssl_key — certificate material is write-only and never leaves the panel again.

Deliberately dropped: automatic node-side provisioning — answering traffic for the domain needs a web server on the node, and the competitor's automation needs a node-side agent; this addon stores/validates the rules and renders a copy-paste nginx server block per rule instead.

Common mistakes: a rule shows "DNS pending" until the domain's A record resolves to the node's public address; raw TCP/UDP game protocols are better served by SRV records (see Subdomain Manager) or an nginx stream block.

---

Server Importer

Sub-tabs: Overview (connections + runs) and Run report. No persistent settings — all inputs are per-operation.

Managed data:

• Connections — name, base_url (public root URL of the source panel), token (Application API key ptla_…, min 16 chars, stored encrypted, with read access to servers, users, nodes, nests and eggs), verify_tls (on; uncheck for self-signed panels).
• Runs (dry run form) — connection_id, target_node_id (allocations claimed from its free pool), fallback_egg_id (optional; unmatched eggs otherwise become conflicts), owner_strategy (map by email / fixed one account), fixed_user_id, create_missing_users, include_suspended, skip_scripts.
• Config-only (config/server-importer.php): per_page 50, timeout 15 s, connect_timeout 8 s, max_servers_per_run 500.

Common mistakes: skipping the dry run — it builds the full conflict report and changes nothing; executing against a node without enough free allocations (claims are re-verified between dry run and import). Rollback force-deletes every server a run created.

---

Server Properties

Single settings pane.

Setting	Type	Default	What it does / when to change
nav_label	text (max 30)	Properties	Label of the tab in the server sub-navigation.
file_name	text (max 64)	server.properties	File in the server root the editor reads/writes. Bare file names only.
allowed_nests	text	1	Comma-separated nest IDs the editor is available for (1 = stock Minecraft). Empty = every nest.
show_descriptions	checkbox	on	Short explanation under each known key.
enable_motd	checkbox	on	MOTD editor (legacy § codes, MiniMessage, raw) with live server-list preview.
enable_icon	checkbox	on	Server icon manager — center-crop to exactly 64×64 PNG, verified again server-side.
enable_yaml	checkbox	on	Auto-detected YAML editor for root-level .yml/.yaml files, with BungeeCord-style proxy support and a raw mode.
max_file_bytes (env-only)	number	1048576 (SERVERPROPERTIES_MAX_FILE_BYTES)	Refuse to read files larger than this.

The four original UI settings also have env defaults (SERVERPROPERTIES_NAV_LABEL, SERVERPROPERTIES_FILE_NAME, SERVERPROPERTIES_ALLOWED_NESTS, SERVERPROPERTIES_SHOW_DESCRIPTIONS).

Common mistakes: expecting instant effect — Minecraft reads server.properties, server-icon.png and the YAML files at boot, so changes apply after a restart; the structured YAML editor rewrites the file from the parsed structure, so YAML comments are not preserved there (use raw mode for hand-tuned files); MiniMessage <rainbow> tags are not rendered in the MOTD preview and show literally.

---

Server Splitter

Sub-tabs: Overview (counters + policy summary), Whitelist, Splits, Link Servers, Settings.

Setting	Type	Default	What it does / when to change
enabled	checkbox	on	Master switch for the client Splitter page.
allow_unlimited	checkbox	on	Unlimited masters (limit 0) may hand out unlimited sub-servers.
whitelist_only	checkbox	off	Only whitelisted servers may split.
whitelist_fallback	checkbox	on	When a server has both a whitelist entry and matching rules, the higher allowance wins.
min_master_cpu / min_master_memory / min_master_disk	numbers	50 / 512 / 1024	Must remain on the master after a split (% / MB / MB).
min_sub_cpu / min_sub_memory / min_sub_disk	numbers	25 / 128 / 512	Every sub-server must receive at least this.
max_ports_per_sub (env-only)	number	10 (SERVER_SPLITTER_MAX_PORTS)	Port cap per sub-server.
sub_description (env-only)	text	Sub-server of {master} (SERVER_SPLITTER_SUB_DESCRIPTION)	Description template for created sub-servers.

Managed data: whitelist entries (per-server fixed limits), resource-graded rules ("at least 1 GB RAM may run 2 sub-servers") checked in an admin-defined priority order, and admin links (master_server_id ↔ sub_server_id) that attach an existing standalone server without moving resources.

Common mistakes: a sub-server stuck in the installing state means the node was unreachable at creation — the bookkeeping is still correct, so reinstall it (or remove the split) once the node answers; after deleting sub-servers outside the splitter, run php artisan server-splitter:reconcile to credit the master back and clear stale markers.

---

Server Type Changer

Sub-tabs: Overview (settings + the last 25 change runs across all servers) and Whitelist (per-server grants by short identifier or UUID, with a name/identifier search helper).

Setting	Type	Default	What it does / when to change
enabled	checkbox	off	Master switch — the client page and API stay off until enabled.
allow_nest_changes	checkbox	on	Permit cross-nest moves.
allow_egg_changes	checkbox	on	Permit in-nest egg swaps.
blocked_nests[] / blocked_eggs[]	multi-selects	empty	Nests/eggs excluded from selection.
event_retention	number (1–200)	25	History entries kept per server.

Managed data: whitelist grants; without one the nav entry stays hidden and the API rejects requests (root admins and the settings.reinstall permission govern access).

Common mistakes: a failed safety backup aborts the whole change before anything is touched — that is the point, check backup capacity instead of retrying blindly; suspended, installing, restoring or transferring servers are rejected until they settle; node-unreachable steps complete as completed with warnings naming the exact step that could not run.

---

Staff Requests

Single pane: policy settings, stat tiles, staff roster management, recent requests (with type + urgent badges and message previews), active grants, audit trail (the three tables are read-only; approve/deny/revoke happens on the user side).

Setting	Type	Default	What it does / when to change
access_mode	select	everyone	Who can use Staff Requests: everyone, allowlist, blacklist. Root admins always bypass.
allowlist	textarea	empty	One email per line; only these accounts may use the feature (shown when mode = allowlist).
blacklist	textarea	empty	One email per line; these accounts are blocked (shown when mode = blacklist).
request_decay_hours	number (1–8760)	24	Unanswered requests expire after this; runs on the panel's schedule cron.
block_rate_limit	number (1–1000)	10	Block-user actions per user per minute.
request_ttl_days (env-only)	number	14 (ASKACCESS_REQUEST_TTL_DAYS)	Separate backend validity window for pending requests — distinct from request_decay_hours.
max_pending_per_user (env-only)	number	25 (ASKACCESS_MAX_PENDING)	Pending-request cap per user.
create_rate_limit / create_rate_window_minutes (env-only)	numbers	15 / 60	Request-creation rate limit.
max_servers_per_grant (env-only)	number	250	Cap on servers touched by a single "all servers" approval.

Managed data (staff roster): email (required — the account is added by email, never enumerated) and an optional role label like "Support". Roster members receive owner-initiated Request Staff Help requests; usernames are shown to users, emails are never exposed, and staff can still opt out via their own accept-requests toggle or blocks.

Common mistakes: confused by the two expiry knobs (request_decay_hours in the UI vs request_ttl_days in env — both exist); expecting uninstall to clean up granted access — approved requests created real subuser rows that survive uninstall.

---

Subdomain Manager

Sub-tabs: Settings, Domains, Records, Limits.

Settings tab (admin.subdomainplus.settings, plus Test Cloudflare connection):

Setting	Type	Default	What it does / when to change
cf_token	password	empty	Cloudflare API token with Zone / DNS / Edit. Placeholder shows "(saved - enter to change)" once set.
cf_email / cf_api_key	email / password	empty	Deprecated Global API Key mode — only use when a token is impossible.
default_quota	number (0–100)	5	DNS records per server; 0 = unlimited.
rate_per_minute	number (1–120)	5	Requests per minute per server.

Managed data (domains): domain and zone_id (required — or click Fetch zones from Cloudflare to pick from your account and auto-fill both), allow_srv_record / allow_a_record (both on), allow_cname_record (on; a requested A record becomes a CNAME when the allocation alias is a hostname — the Use allocation IP checkbox keeps the old behavior), target_ip (optional forced IP; empty auto-resolves from each server's allocation FQDN), SRV presets (service default _minecraft, proto _tcp/_udp, priority 0, weight 5, ttl 1 = automatic), banned_patterns (PHP regex, one per line), allowed_eggs[] (empty = all).

Managed data (Limits tab): per-server quota overrides (server_id + quota) and the node policy — node_policy (allow or deny) with node_ids[]; restricted servers keep viewing/deleting existing records but cannot create new ones.

Common mistakes: a token without Zone / DNS / Edit scope; expecting proxied records — the addon always creates unproxied records (SRV and CNAME require it); missing the panel cron, which drives subdomainplus:sync-all and orphan cleanup. The create form's live availability check blocks known-taken names before submit.

---

Version Changer

Settings pane plus a recent-changes table (saving settings also flushes the manifest cache), and a Server Access sub-tab for per-server grants (by short id or UUID).

Setting	Type	Default	What it does / when to change
enabled	checkbox	on	Off hides the picker in the client panel and rejects installs.
allowed_software[]	checkboxes	all	Software entries users may see: vanilla, paper, purpur, folia, velocity, waterfall.
allowed_types[]	checkboxes	release only	Vanilla version types users may install: release, snapshot, old_beta, old_alpha.
allow_wipe	checkbox	on	Permit the archive-then-wipe and plain-wipe data options per install.
require_grant	checkbox	off	Restrict the feature to servers on the Server Access list.
default_filename	text	server.jar	Target jar name; must match a plain .jar file name (validated, resets to server.jar otherwise).
cache_ttl	number (60–86400)	300	Seconds the version/build catalogues are cached.
history_limit	number (0–500)	50	History entries kept per server, pruned after each install.
allow_filename_choice	checkbox	on	Let users override the jar name per install.
backup_previous	checkbox	on	Rename the current jar to <name>.backup-<timestamp> before writing (keep-files installs).
update_startup_variable	checkbox	on	Point the egg's SERVER_JARFILE variable at the new jar.
manifest_url / http_timeout / http_retries (env-only)	—	piston-meta URL / 10 / 2	Vanilla version-service endpoint and HTTP tuning (VERSION_CHANGER_MANIFEST_URL, VERSION_CHANGER_HTTP_TIMEOUT, VERSION_CHANGER_HTTP_RETRIES).

Common mistakes: assuming a mirror manifest widens download sources — jar downloads are pinned to the official hosts of each software (mojang.com / minecraft.net, papermc.io, purpurmc.org) regardless of the catalogue endpoints; expecting the running server to switch jars live — the old jar runs until the next restart.

---

Votifier Tester

Single pane: global toggle, default port and timeouts.

Setting	Type	Default	What it does / when to change
enabled	checkbox	on	Off disables the client page and API panel-wide.
default_port	number (1–65535)	8192	Pre-filled listener port on the test form.
connect_timeout	number (1–15)	5.0 s	TCP connect timeout.
read_timeout	number (1–30)	5.0 s	Response timeout.
service_name (env-only)	text	VotifierTest (VOTIFIER_TESTER_SERVICE_NAME)	Service name sent in test votes.

Common mistakes: the listener port must be reachable from the panel host (open the firewall / publish it for remote nodes); a well-formed vote is answered with silence — a clean handshake plus transmission counts as a pass, and the reward event shows up in the server console; tokens and public keys travel with the test only and are never persisted.

---

World Manager

Single settings pane.

Setting	Type	Default	What it does / when to change
curseforge_api_key	text	empty	From the CurseForge console; the catalogue stays unavailable to users until one is saved.
default_sort	select	downloads	Preselected catalogue ordering.
nav_label	text (max 32)	Worlds	Client tab label.
catalog_ttl	number (60–86400)	900	Catalogue cache lifetime in seconds.
http_timeout / version_manifest_ttl / progress_ttl (env-only)	seconds	15 / 86400 / 600 (MC_WORLD_MANAGER_HTTP_TIMEOUT, …_MANIFEST_TTL, …_PROGRESS_TTL)	Request timeout, MC version manifest cache, install progress record lifetime.

Common mistakes: archives Wings cannot unpack (rar, 7z, plain gzip) are rejected with a clear message — repack as zip; switching the active world rewrites level-name in server.properties, so restart to play; manually uploaded worlds under a different directory name simply show without a catalogue badge.

---

What's Next?

• The Hub Dashboard → — how panes, sub-tabs and saves work.
• Addons Guide → — per-addon usage notes and client surfaces.