RELAXY! - CLUSTER MANAGER CONSOLE ================================= Type a command at the manager prompt. "show help" prints this list. Two things are deliberately separate and are often confused: the OPERATING MODE what Relaxy tells users it is doing (normal / debug / maintenance / silent). Changed ONLY with "update status". It persists in assets/config/status.ini. the LOG CHANNELS what gets printed to your console. Changed with "toggle". Purely local noise control; users never see it. Turning logging on does NOT put the bot into maintenance, and vice versa. PROCESS CONTROL --------------- - halt "on" stops every cluster and disables auto-respawn. "off" respawns the clusters and re-enables auto-respawn. - restart shard Restarts a single cluster process. - restart module manager Hot-reloads one manager-level module, or all of them. - restart module self_repair Restarts the file watcher / self-repair mechanism. - restart module cluster Broadcasts a hot-reload of the named module to every cluster. - recluster [delay_ms] Reclusters all shards, optionally after a delay (default 0). - sync [force] "git pull" on the bot's workspace. "force" runs "git reset --hard" first. - ping Latency between the console input and the manager's response. - eval Runs JavaScript inside the MANAGER process. Returns the result or the error. USE WITH EXTREME CAUTION - this is the manager, not a shard. INSPECTION - "show " -------------------------------- - show help This list. - show shards Detailed state of every shard/cluster. - show shard The raw internal cluster iterator. - show pids PIDs of the manager and of every live cluster. - show uptime How long the manager has been up. - show opmode Current operating mode (NORMAL / DEBUG / MAINTENANCE / ...). - show version The version string currently advertised. - show trueshard Which cluster is the designated "true Relaxy" cluster. - show modules Loaded manager modules and their file paths. - show owners Asks each cluster to report the configured owner IDs. - show servers Server count per cluster, and the fleet total. - show queues Voice connections and queue depth, from any cluster that has one. - show dependencies Asks cluster #0 to report its music dependency versions. - show cpusage CPU usage for one cluster, or all of them. - show heartbeat [all | clusterID] Heartbeat stats: total, missed, average latency, suffocate flag. - show watchdog Internet watchdog: recovery state, last restart, cooldown, and per-shard error counts. - show logs Every log channel, whether it is ON or OFF, which process prints it, and what it covers. See LOGGING below. LOGGING - "toggle " ---------------------------- Every category of logging is a named CHANNEL. Toggling one flips it on the manager AND on every cluster at once, so the fleet stays in step; a cluster that respawns later comes back with the toggles you set, not with the defaults. toggle Flip a single channel. toggle debug MASTER SWITCH - flips EVERY channel at once. (If any channel is off, this turns them all on; if they are all on, it turns them all off.) "toggle debug" does NOT change the operating mode. Use "update status" for that. Channels: debug General debug output (and the master switch, above) heartbeats Cluster heartbeat ticks inter_process_messages IPC traffic between the manager and its clusters file_operations File handler reads/writes garbage_collector Cleanup of deleted Discord entities (channels, roles, guilds) cpu_monitor Periodic CPU usage sampling cpu_monitor_file_logs Also write the CPU profiles to disk music_debug Music player / extractor debug output raw_events Raw gateway events, before dispatch event_names The name of every dispatched event rate_limiter Rate limiter queueing and releases message_debug Per-message command pipeline tracing (owner messages only) database Database queue flushes and worker writes Run "show logs" to see the current state of all of them. Other toggles (not log channels): toggle low_performance Low-performance mode (gates the resource-heavy features). toggle output Pause / resume the manager's own stdout. toggle usage_reports Periodic usage reports from the clusters on/off. CONFIGURATION - "update " ------------------------------------- These persist in the "settings" collection in MongoDB, so they survive restarts and stay consistent across every host. They are NOT read from .env at runtime. - update status Sets the OPERATING MODE - what Relaxy advertises to users. normal Business as usual. debug Being worked on, still usable. Also enables debug logging. maintain Maintenance. Also enables debug logging. silent Running, but not talking. exit Begins a graceful shutdown of the whole fleet. - update version Sets the version string, saves it, and broadcasts it to every cluster. - update owners [id, id, ...] With no arguments, prints the current owner IDs. Otherwise replaces the owner list (comma or space separated Discord snowflakes). Owners are a LIST; the first one is the primary owner (the account DM'd for requests and version reports). - update cookie [ | clear] The shared YouTube cookie used by the music module, fleet-wide. With no arguments, reports whether one is set and how many cookies it holds - the value itself is a credential and is never printed. Pass a Netscape-format cookies.txt to replace it, or "clear" to remove it. - update usage_reports How often the clusters send their periodic usage reports. SLASH COMMANDS -------------- - create slash Tells every cluster to register the slash commands. - delete slash Tells every cluster to delete all of its slash commands. Note: only the fun, image, miscellaneous, moderation and music categories are ever registered. Commands in "administrator" are owner-only dev tools and are deliberately excluded (as is anything marked leaveout).