fix(perf): improve delete_from_sorted_set speed

lib/sidekiq_unique_jobs/lua/shared/_delete_from_sorted_set.lua

@@ -1,19 +1,22 @@
 local function delete_from_sorted_set(name, digest)
-  local per   = 50
-  local total = redis.call("zcard", name)
-  local index = 0
-  local result
+  local cursor = "0"
+  local result = nil
+  local match_pattern = "*" .. digest .. "*"
 
-  while (index < total) do
-    local items = redis.call("ZRANGE", name, index, index + per -1)
-    for _, item in pairs(items) do
+  repeat
+    local scan_result = redis.call("ZSCAN", name, cursor, "MATCH", match_pattern, "COUNT", 1)
+    cursor = scan_result[1]
+    local items = scan_result[2]
+
+    for i = 1, #items, 2 do
+      local item = items[i]
       if string.find(item, digest) then
         redis.call("ZREM", name, item)
         result = item
         break
       end
     end
-    index = index + per
-  end
+  until cursor == "0" or result ~= nil
+
   return result
 end

spec/performance/unique_job_on_conflict_replace_spec.rb

@@ -0,0 +1,57 @@
+# frozen_string_literal: true
+
+RSpec.describe UniqueJobOnConflictReplace, :perf do
+  let(:lock_prefix)  { SidekiqUniqueJobs.config.lock_prefix }
+  let(:lock_timeout) { SidekiqUniqueJobs.config.lock_timeout }
+  let(:lock_ttl)     { SidekiqUniqueJobs.config.lock_ttl }
+  let(:queue)        { described_class.sidekiq_options["queue"] }
+  let(:on_conflict)  { described_class.sidekiq_options["on_conflict"] }
+  let(:lock)         { described_class.sidekiq_options["lock"] }
+
+  before do
+    digests.delete_by_pattern("*")
+  end
+
+  context "when schedule queue is large" do
+    it "locks and replaces quickly" do
+      (0..100_000).each_slice(1_000) do |nums|
+        redis do |conn|
+          conn.pipelined do |pipeline|
+            nums.each do |num|
+              created_at   = Time.now.to_f
+              scheduled_at = created_at + rand(3_600..2_592_000)
+
+              payload = {
+                "retry" => true,
+                "queue" => queue,
+                "lock" => lock,
+                "on_conflict" => on_conflict,
+                "class" => described_class.name,
+                "args" => [num, { "type" => "extremely unique" }],
+                "jid" => SecureRandom.hex(12),
+                "created_at" => created_at,
+                "lock_timeout" => lock_timeout,
+                "lock_ttl" => lock_ttl,
+                "lock_prefix" => lock_prefix,
+                "lock_args" => [num, { "type" => "extremely unique" }],
+                "lock_digest" => "#{lock_prefix}:#{SecureRandom.hex}"
+              }
+
+              pipeline.zadd("schedule", scheduled_at, payload.to_json)
+            end
+          end
+        end
+      end
+
+      # queueing it once at the end of the queue should succeed
+      expect(described_class.perform_in(2_592_000, 100_000, { "type" => "extremely unique" })).not_to be_nil
+
+      # queueing it again should be performant
+      expect do
+        Timeout.timeout(0.1) do
+          described_class.perform_in(2_592_000, 100_000, { "type" => "extremely unique" })
+        end
+      end.not_to raise_error
+    end
+  end
+end

spec/support/workers/unique_job_on_conflict_replace.rb

@@ -0,0 +1,14 @@
+# frozen_string_literal: true
+
+# :nocov:
+
+class UniqueJobOnConflictReplace
+  include Sidekiq::Worker
+  sidekiq_options lock: :until_executing,
+                  queue: :customqueue,
+                  on_conflict: :replace
+
+  def perform(one, two)
+    [one, two]
+  end
+end

spec/workers/unique_job_on_conflict_replace_spec.rb

@@ -0,0 +1,18 @@
+# frozen_string_literal: true
+
+RSpec.describe UniqueJobOnConflictReplace do
+  it_behaves_like "sidekiq with options" do
+    let(:options) do
+      {
+        "lock" => :until_executing,
+        "on_conflict" => :replace,
+        "queue" => :customqueue,
+        "retry" => true,
+      }
+    end
+  end
+
+  it_behaves_like "a performing worker" do
+    let(:args) { ["hundred", { "type" => "extremely unique", "id" => 44 }] }
+  end
+end