Session resumption

Lavalink can keep a session alive for a configurable window after your bot disconnects. If you reconnect within that window and present the original session ID, Lavalink resumes — queues keep playing, no interruption.

Discolink doesn’t implement this automatically. The right approach depends on your scale and persistence layer.

1. Opt in and persist the session ID

   On a fresh connection, tell Lavalink to keep the session alive and store the session ID:

   player.on("nodeReady", async (node, resumed) => {
     if (!resumed) {
       // Enable resumption with a 60-second window
       await node.rest.updateSession({ resuming: true, timeout: 60 });
       // Persist the session ID — use your preferred storage
       await db.set(`session:${node.name}`, node.sessionId);
     }
   });

2. Pass the session ID on reconnect

   Disable auto init and initialize nodes with their session IDs:

   const player = new Player({
     autoInit: false,
     // don't pass your nodes here anymore
     // ...
   });
   
   // Retrieve the stored session ID, then initialize
   const storedSessionId = await db.get("session:main");
   
   await player.init(clientId, [
     {
       name: "main",
       origin: "http://localhost:2333",
       password: "youshallnotpass",
       sessionId: storedSessionId ?? undefined,
     },
   ]);

3. Sync local state after a successful resume

   After a node reconnects, update your queues:

   player.on("nodeReady", async (node, resumed) => {
     if (!resumed) {
       // Clear data associated with the node
       await db.del(`session:${node.name}`);
       await db.del(`queues:${node.name}`);
       return;
     }
   
     // Re-create queues from saved data
     const queues = await db.get(`queues:${node.name}`);
     for (const q of unpackQueues(queues)) await player.queues.create(q);
   
     // Note: syncing does not add a (current) track
     // Note: if there are too many queues, sync them individually not in bulk
     await player.queues.sync(node);
   });