fix: ReplicationConnection init timeout

[edgurgel]

What kind of change does this PR introduce?

Add a GenServer wrapper that we can control a timeout to initialise.

What is the current behavior?

We think that Postgrex.ReplicationConnection with sync_connect: true can block indefinitely.

Here is one of the partitioned supervisors blocking on Postgrex.ReplicationConnection.start_link:

realtime/lib/realtime/tenants/replication_connection.ex

Line 124 in 44c0543

case Postgrex.ReplicationConnection.start_link(__MODULE__, attrs, connection_opts) do

iex> Process.info(pid, :current_stacktrace)
{:current_stacktrace,
 [
   {:proc_lib, :sync_start, 2, [file: ~c"proc_lib.erl", line: 434]},
   {Realtime.Tenants.ReplicationConnection, :start_link, 1,
    [file: ~c"lib/realtime/tenants/replication_connection.ex", line: 124]},
   {DynamicSupervisor, :start_child, 3,
    [file: ~c"lib/dynamic_supervisor.ex", line: 795]},
   {DynamicSupervisor, :handle_start_child, 2,
    [file: ~c"lib/dynamic_supervisor.ex", line: 781]},
   {:gen_server, :try_handle_call, 4, [file: ~c"gen_server.erl", line: 2381]},
   {:gen_server, :handle_msg, 6, [file: ~c"gen_server.erl", line: 2410]},
   {:proc_lib, :init_p_do_apply, 3, [file: ~c"proc_lib.erl", line: 329]}
 ]}

This supervisor was stuck for a very long time.

Notice that Realtime.Tenants.ReplicationConnection.init/1 does not do any lengthy work that could block Postgrex.ReplicationConnection.init/1

realtime/lib/realtime/tenants/replication_connection.ex

Lines 133 to 147 in 44c0543

	def init(%__MODULE__{tenant_id: tenant_id, monitored_pid: monitored_pid} = state) do
	Logger.metadata(external_id: tenant_id, project: tenant_id)
	Process.monitor(monitored_pid)
	state = %{state | table: "messages", schema: "realtime"}
	state = %{
	state
	| publication_name: publication_name(state),
	replication_slot_name: replication_slot_name(state)
	}
	Logger.info("Initializing connection with the status: #{inspect(state, pretty: true)}")
	{:ok, state}
	end

The ideal solution here would be to change Postgrex.ReplicationConnection.start_link to accept a new timeout option for the init step https://github.com/elixir-ecto/postgrex/blob/257daa773a7558d574df3aa3b558664275787ff8/lib/postgrex/replication_connection.ex#L356-L366

What is the new behavior?

It can't block because the GenServer wrapper will eventually timeout on init causing them to stop

Additional context

Add any other context or screenshots.

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

1 Skipped Deployment

Name	Status	Preview	Comments	Updated (UTC)
realtime-demo	⬜️ Ignored (Inspect)			Jun 23, 2025 1:31am

[coveralls]

coverage: 84.89% (+0.4%) from 84.505%
when pulling b186a32 on fix/postgrex-replication-init-timeout
into 44c0543 on main.

[kiwicopple]

🎉 This PR is included in version 2.37.4 🎉

The release is available on GitHub release

Your semantic-release bot 📦🚀