Close a KubeMQ Ruby Client

Overview

Closing a client isn't an afterthought — it tells the broker and your own process that this connection is done, so both sides release what they were holding for it. A KubeMQ client is more than a socket: it's a gRPC channel plus whatever subscriptions and buffered messages it's servicing. Skip the close and those linger — subscriptions keep streaming, the channel stays open — and in short-lived scripts or test suites you leak connections until the process exits.

Calling close cancels active subscriptions, flushes the message buffer, and closes the underlying gRPC channel. Once it returns, closed? flips to true for the rest of the client's lifetime.

Gotchas: the drain window is bounded, not unlimited, so a slow consumer can still lose the tail of a burst if you close mid-stream; a closed client is dead forever — no reconnect on the same instance, build a new one; and close is idempotent, so calling it a second time is safe, but any other method call after closing raises ClientClosedError.

Prerequisites

• KubeMQ server running on localhost:50000
• Ruby SDK installed (gem install kubemq)

Code

require 'kubemq'

address = ENV.fetch('KUBEMQ_ADDRESS', 'localhost:50000')

begin
  client = KubeMQ::PubSubClient.new(address: address, client_id: 'close-example')
  puts "Connected to #{address}"

  info = client.ping
  puts "Ping OK: host=#{info.host}, version=#{info.version}"

  client.close
  puts "Client closed (closed?=#{client.closed?})"

  client.close
  puts 'Second close is safe (idempotent)'
rescue KubeMQ::Error => e
  puts "KubeMQ error: #{e.message}"
end

How It Works

• close cancels active subscriptions, flushes the message buffer, and closes the gRPC channel.
• Calling close multiple times is safe — the method is idempotent.
• After closing, closed? returns true and further operations raise ClientClosedError.

Related

• Ruby SDK Reference
• Connect
• Graceful Shutdown