From 92544e8f99658be53161caa8c4b5498ae2271d10 Mon Sep 17 00:00:00 2001
From: Oneric <oneric@oneric.stub>
Date: Tue, 29 Oct 2024 01:41:54 +0100
Subject: [PATCH] Don't enqueue a plethora of unnecessary NodeInfoFetcher jobs

There were two issues leading to needles effort:
Most importnatly, the use of AP IDs as "source_url" meant multiple
simultaneous jobs got scheduled for the same instance even with the
default unique settings.
Also jobs were scheduled uncontionally for each processed AP object
meaning we incured oberhead from managing Oban jobs even if we knew it
wasn't necessary. By comparison the single query to check if an update
is needed should be cheaper overall.
---
 lib/pleroma/instances/instance.ex             |  8 +++++++
 .../workers/nodeinfo_fetcher_worker.ex        | 23 ++++++++++++++++++-
 .../web/activity_pub/side_effects_test.exs    |  2 +-
 3 files changed, 31 insertions(+), 2 deletions(-)

diff --git a/lib/pleroma/instances/instance.ex b/lib/pleroma/instances/instance.ex
index 5c70748b6..63362eb28 100644
--- a/lib/pleroma/instances/instance.ex
+++ b/lib/pleroma/instances/instance.ex
@@ -158,6 +158,14 @@ def needs_update(%Instance{metadata_updated_at: metadata_updated_at}) do
     NaiveDateTime.diff(now, metadata_updated_at) > 86_400
   end
 
+  def needs_update(%URI{host: host}) do
+    with %Instance{} = instance <- Repo.get_by(Instance, %{host: host}) do
+      needs_update(instance)
+    else
+      _ -> true
+    end
+  end
+
   def local do
     %Instance{
       host: Pleroma.Web.Endpoint.host(),
diff --git a/lib/pleroma/workers/nodeinfo_fetcher_worker.ex b/lib/pleroma/workers/nodeinfo_fetcher_worker.ex
index 27492e1e3..32907bac9 100644
--- a/lib/pleroma/workers/nodeinfo_fetcher_worker.ex
+++ b/lib/pleroma/workers/nodeinfo_fetcher_worker.ex
@@ -1,9 +1,30 @@
 defmodule Pleroma.Workers.NodeInfoFetcherWorker do
-  use Pleroma.Workers.WorkerHelper, queue: "nodeinfo_fetcher"
+  use Pleroma.Workers.WorkerHelper,
+    queue: "nodeinfo_fetcher",
+    unique: [
+      keys: [:op, :source_url],
+      # old jobs still get pruned after a short while
+      period: :infinity,
+      states: Oban.Job.states()
+    ]
 
   alias Oban.Job
   alias Pleroma.Instances.Instance
 
+  def enqueue(op, %{"source_url" => ap_id} = params, worker_args) do
+    # reduce to base url to avoid enqueueing unneccessary duplicates
+    domain =
+      ap_id
+      |> URI.parse()
+      |> URI.merge("/")
+
+    if Instance.needs_update(domain) do
+      do_enqueue(op, %{params | "source_url" => URI.to_string(domain)}, worker_args)
+    else
+      :ok
+    end
+  end
+
   @impl Oban.Worker
   def perform(%Job{
         args: %{"op" => "process", "source_url" => domain}
diff --git a/test/pleroma/web/activity_pub/side_effects_test.exs b/test/pleroma/web/activity_pub/side_effects_test.exs
index 28a591d3c..64a1fe6e6 100644
--- a/test/pleroma/web/activity_pub/side_effects_test.exs
+++ b/test/pleroma/web/activity_pub/side_effects_test.exs
@@ -46,7 +46,7 @@ test "it queues a fetch of instance information" do
 
       assert_enqueued(
         worker: Pleroma.Workers.NodeInfoFetcherWorker,
-        args: %{"op" => "process", "source_url" => "https://wowee.example.com/users/1"}
+        args: %{"op" => "process", "source_url" => "https://wowee.example.com/"}
       )
     end
   end