On this page

Extends: undici.Dispatcher

A pool of Client instances connected to the same upstream target with round-robin client selection.

Unlike Pool, which always selects the first available client, RoundRobinPool cycles through clients in a round-robin fashion. This ensures even distribution of requests across all connections, which is particularly useful when the upstream target is behind a load balancer that round-robins TCP connections across multiple backend servers (e.g., Kubernetes Services).

Requests are not guaranteed to be dispatched in order of invocation.

new RoundRobinPool(url, options?): void

Arguments:

  • url URL | string - It should only include the protocol, hostname, and port.
  • options RoundRobinPoolOptions (optional)

Extends: ClientOptions

  • factory (origin: URL, opts: Object) => Dispatcher - Default: (origin, opts) => new Client(origin, opts)
  • connections number | null (optional) - Default: null - The number of Client instances to create. When set to null, the RoundRobinPool instance will create an unlimited amount of Client instances.
  • clientTtl number | null (optional) - Default: null - The amount of time before a Client instance is removed from the RoundRobinPool and closed. When set to null, Client instances will not be removed or closed based on age.

RoundRobinPool is designed for scenarios where:

  1. You connect to a single origin (e.g., http://my-service.namespace.svc)
  2. That origin is backed by a load balancer distributing TCP connections across multiple servers
  3. You want requests evenly distributed across all backend servers

Example: In Kubernetes, when using a Service DNS name with multiple Pod replicas, kube-proxy load balances TCP connections. RoundRobinPool ensures each connection (and thus each Pod) receives an equal share of requests.

RoundRobinPool distributes HTTP requests evenly across TCP connections. Whether this translates to even backend server distribution depends on the load balancer's behavior:

✓ Works when the load balancer:

  • Assigns different backends to different TCP connections from the same client
  • Uses algorithms like: round-robin, random, least-connections (without client affinity)
  • Example: Default Kubernetes Services without sessionAffinity

✗ Does NOT work when:

  • Load balancer has client/source IP affinity (all connections from one IP → same backend)
  • Load balancer uses source-IP-hash or sticky sessions

How it works:

  1. RoundRobinPool creates N TCP connections to the load balancer endpoint
  2. Load balancer assigns each TCP connection to a backend (per its algorithm)
  3. RoundRobinPool cycles HTTP requests across those N connections
  4. Result: Requests distributed proportionally to how the LB distributed the connections

If the load balancer assigns all connections to the same backend (e.g., due to session affinity), RoundRobinPool cannot overcome this. In such cases, consider using BalancedPool with direct backend addresses (e.g., individual pod IPs) instead of a load-balanced endpoint.

Implements Client.closed

Implements Client.destroyed

Returns PoolStats instance for this pool.

RoundRobinPool.close(callback?): void

Implements Dispatcher.close([callback]).

RoundRobinPool.destroy(error?, callback?): void

Implements Dispatcher.destroy([error, callback]).

RoundRobinPool.connect(options, callback?): void

See Dispatcher.connect(options[, callback]).

RoundRobinPool.dispatch(options, handler): void

Implements Dispatcher.dispatch(options, handler).

RoundRobinPool.pipeline(options, handler): void

See Dispatcher.pipeline(options, handler).

RoundRobinPool.request(options, callback?): void

See Dispatcher.request(options [, callback]).

RoundRobinPool.stream(options, factory, callback?): void

See Dispatcher.stream(options, factory[, callback]).

RoundRobinPool.upgrade(options, callback?): void

See Dispatcher.upgrade(options[, callback]).

See Dispatcher Event: 'connect'.

See Dispatcher Event: 'disconnect'.

See Dispatcher Event: 'drain'.

import { RoundRobinPool } from 'undici'

const pool = new RoundRobinPool('http://my-service.default.svc.cluster.local', {
  connections: 10
})

// Requests will be distributed evenly across all 10 connections
for (let i = 0; i < 100; i++) {
  const { body } = await pool.request({
    path: '/api/data',
    method: 'GET'
  })
  console.log(await body.json())
}

await pool.close()
JavaScript
  • Pool - Connection pool without round-robin
  • BalancedPool - Load balancing across multiple origins
  • Issue #3648 - Original issue describing uneven distribution