@DiscordJS/WS cache

By default, does djs/ws cache anything? for context I have GuildPresences, Guilds, and GuildMembers intents enabled.

19 Replies

- What's your exact discord.js npm list discord.js and node node -v version? - Not a discord.js issue? Check out #other-js-ts. - Consider reading #how-to-get-help to improve your question! - Explain what exactly your issue is. - Post the full error stack trace, not just the top part! - Show your code! - Issue solved? Press the button! - ✅ Marked as resolved by OP

JaymartOP•6mo ago

I would assume not seeing as it's a barebones but I'm somehow running at like 20 GBs or more

Unknown User•6mo ago

Message Not Public

JaymartOP•6mo ago

20k guilds, last I checked was somewhere around 20 million users im currently rewriting it from python lol, so the bot itself is very miniscule so im not sure why it's eating up so much ram unless there is stuff being cached by default

JaymartOP•6mo ago

like.

JaymartOP•6mo ago

all its doing is receiving presence update and pushing it to rabbitMQ thats it unless somehow the data im pushing is staying within this instance? for more context

import { Broker } from '@vanityroles/broker';
import { WebSocketManager } from '@discordjs/ws';
import { GatewayDispatchEvents, GatewayIntentBits, Client } from '@discordjs/core';
import { REST } from '@discordjs/rest';
import { createClient } from 'redis';
import dotenv from 'dotenv';

dotenv.config();

// Set up your REST client
const rest = new REST({ version: '10' }).setToken(`${process.env.DISCORD_TOKEN}`);

// Initialize your broker
const broker = new Broker();

// Create Redis client
const redisClient = createClient({ url: process.env.REDIS_PRIVATE_URL });

const connectRedis = async () => {
  try {
    if (!redisClient.isOpen) {
      await redisClient.connect();
    }
  } catch (error) {
    console.error('Failed to connect to Redis:', error);
  }
};

const getClusterId = async (): Promise<number> => {
  await connectRedis();
  const clusterId = await redisClient.incr('cluster_id_counter');
  return clusterId - 1; // Adjusting for zero-based indexing
};

const getShardIds = (totalShards: number, clusterId: number, clusters: number): number[] => {
  const shardsPerCluster = Math.ceil(totalShards / clusters);
  const shardIds = [];

  for (let i = 0; i < shardsPerCluster; i++) {
    const shardId = clusterId * shardsPerCluster + i;
    if (shardId < totalShards) {
      shardIds.push(shardId);
    }
  }

  return shardIds;
};

const delay = (ms: number) => new Promise(resolve => setTimeout(resolve, ms));

const main = async () => {
  try {
    await connectRedis();

    const totalShards = Number(process.env.TOTAL_SHARDS) || 15;
    const totalClusters = Number(process.env.TOTAL_CLUSTERS) || 3;

    // Get cluster ID from Redis
    const clusterId = await getClusterId();

    const shardIds = getShardIds(totalShards, clusterId, totalClusters);
    const shardCount = totalShards;

    console.log(`Cluster ID: ${clusterId}. Shard IDs: ${shardIds}. Shard Count: ${shardCount}`)

    // Calculate the delay based on the clusterId
    const delayInterval = Number(process.env.DELAY_INTERVAL_MS) || 5000; // Default interval of 5 seconds
    const delayMs = clusterId * delayInterval;
    console.log(`Cluster ID: ${clusterId}. Waiting for ${delayMs}ms before starting the gateway...`);
    await delay(delayMs);

    // Configure the gateway with sharding information
    const gateway = new WebSocketManager({
      token: `${process.env.DISCORD_TOKEN}`,
      intents: GatewayIntentBits.GuildPresences,
      rest,
      shardCount,
      shardIds,
    });

    // Initialize the client
    const client = new Client({ rest, gateway });

    // Event handler for presence updates
    client.on(GatewayDispatchEvents.PresenceUpdate, async ({ data: interaction }) => {
      try {
        await broker.sendToQueue('presence', { data: interaction });
      } catch (error) {
        console.error("Failed to handle message: ", error);
      }
    });

    // Event handler for when the bot is ready
    client.once(GatewayDispatchEvents.Ready, () => console.log("Ready!"));

    // Connect to the gateway
    await gateway.connect();
    console.log("Gateway connected");
  } catch (error) {
    console.error('Failed to start bot:', error);
  } finally {
    // Close Redis client only after all operations are done
    await redisClient.quit();
  }
};

void main();

import { Broker } from '@vanityroles/broker';
import { WebSocketManager } from '@discordjs/ws';
import { GatewayDispatchEvents, GatewayIntentBits, Client } from '@discordjs/core';
import { REST } from '@discordjs/rest';
import { createClient } from 'redis';
import dotenv from 'dotenv';

dotenv.config();

// Set up your REST client
const rest = new REST({ version: '10' }).setToken(`${process.env.DISCORD_TOKEN}`);

// Initialize your broker
const broker = new Broker();

// Create Redis client
const redisClient = createClient({ url: process.env.REDIS_PRIVATE_URL });

const connectRedis = async () => {
  try {
    if (!redisClient.isOpen) {
      await redisClient.connect();
    }
  } catch (error) {
    console.error('Failed to connect to Redis:', error);
  }
};

const getClusterId = async (): Promise<number> => {
  await connectRedis();
  const clusterId = await redisClient.incr('cluster_id_counter');
  return clusterId - 1; // Adjusting for zero-based indexing
};

const getShardIds = (totalShards: number, clusterId: number, clusters: number): number[] => {
  const shardsPerCluster = Math.ceil(totalShards / clusters);
  const shardIds = [];

  for (let i = 0; i < shardsPerCluster; i++) {
    const shardId = clusterId * shardsPerCluster + i;
    if (shardId < totalShards) {
      shardIds.push(shardId);
    }
  }

  return shardIds;
};

const delay = (ms: number) => new Promise(resolve => setTimeout(resolve, ms));

const main = async () => {
  try {
    await connectRedis();

    const totalShards = Number(process.env.TOTAL_SHARDS) || 15;
    const totalClusters = Number(process.env.TOTAL_CLUSTERS) || 3;

    // Get cluster ID from Redis
    const clusterId = await getClusterId();

    const shardIds = getShardIds(totalShards, clusterId, totalClusters);
    const shardCount = totalShards;

    console.log(`Cluster ID: ${clusterId}. Shard IDs: ${shardIds}. Shard Count: ${shardCount}`)

    // Calculate the delay based on the clusterId
    const delayInterval = Number(process.env.DELAY_INTERVAL_MS) || 5000; // Default interval of 5 seconds
    const delayMs = clusterId * delayInterval;
    console.log(`Cluster ID: ${clusterId}. Waiting for ${delayMs}ms before starting the gateway...`);
    await delay(delayMs);

    // Configure the gateway with sharding information
    const gateway = new WebSocketManager({
      token: `${process.env.DISCORD_TOKEN}`,
      intents: GatewayIntentBits.GuildPresences,
      rest,
      shardCount,
      shardIds,
    });

    // Initialize the client
    const client = new Client({ rest, gateway });

    // Event handler for presence updates
    client.on(GatewayDispatchEvents.PresenceUpdate, async ({ data: interaction }) => {
      try {
        await broker.sendToQueue('presence', { data: interaction });
      } catch (error) {
        console.error("Failed to handle message: ", error);
      }
    });

    // Event handler for when the bot is ready
    client.once(GatewayDispatchEvents.Ready, () => console.log("Ready!"));

    // Connect to the gateway
    await gateway.connect();
    console.log("Gateway connected");
  } catch (error) {
    console.error('Failed to start bot:', error);
  } finally {
    // Close Redis client only after all operations are done
    await redisClient.quit();
  }
};

void main();

DD•6mo ago

nah, this is on you most likely you're not acking the rmq messages or smth idk by default messages stay in queues until consumed and acked, yes also side note but the way you're handling delays could be better use buildIdentifyThrottler and write a custom one

JaymartOP•6mo ago

they are tho according to the UI.

DD•6mo ago

what is that mem chart you showed is it server wide usage

JaymartOP•6mo ago

railway and no container only

DD•6mo ago

can't really tell then there's an off chance /ws has some sort of memory leak but I doubt it it does not cache anything

JaymartOP•6mo ago

yeah that's what I figured ima disable the event and see what happens

DD•6mo ago

there's bots bigger than yours using it that run fine

JaymartOP•6mo ago

yeee haha was ab to say i hope im not doing that lmao i mean the UI shows them being consumed fast enough but is it possible that in real time they aren't actually? and that's causing the build up of mem?

JaymartOP•6mo ago

i say that because the massive dips in memory

JaymartOP•6mo ago

at one point in there it went from 8gb and jumped to 25 and it does that until it just hits mem limit

JaymartOP•6mo ago

so I was logging as well to see if it matched up.

JaymartOP•6mo ago

it's total used between the 3 instances okay, i disabled the broker/queue and it is indeed the issue. I dont think there are enough consumers even tho the UI is reporting so

JaymartOP•6mo ago

Gaming

Programming

@DiscordJS/WS cache