feat: introduce sidekiq_job_timeout & sidekiq_batch_size

lib/miteru/config.rb

@@ -28,6 +28,8 @@ class Config < Anyway::Config
       sentry_trace_sample_rate: 0.25,
       sidekiq_redis_url: nil,
       sidekiq_job_retry: 0,
+      sidekiq_batch_size: 50,
+      sidekiq_job_timeout: 600,
       cache_redis_url: nil,
       cache_ex: nil,
       cache_prefix: "miteru:cache",
@@ -53,6 +55,12 @@ class Config < Anyway::Config
     # @!attribute [r] sidekiq_job_retry
     #   @return [Integer]
 
+    # @!attribute [r] sidekiq_batch_size
+    #   @return [Integer]
+
+    # @!attribute [r] sidekiq_job_timeout
+    #   @return [Integer]
+
     # @!attribute [r] cache_redis_url
     #   @return [String, nil]
 

lib/miteru/orchestrator.rb

@@ -13,10 +13,8 @@ def call
     end
 
     def sidekiq_call
-      non_cached_websites.each do |website|
-        Jobs::CrawleJob.perform_async(website.url, website.source)
-        logger.info("Website:#{website.truncated_url} crawler job queued.") if verbose?
-      end
+      array_of_args = non_cached_websites.map { |website| [website.url, website.source] }
+      Jobs::CrawleJob.perform_bulk(array_of_args, batch_size: Miteru.config.sidekiq_batch_size)
     end
 
     def parallel_call

lib/miteru/sidekiq/application.rb

@@ -6,7 +6,10 @@
 
 Sidekiq.configure_server do |config|
   config.redis = {url: Miteru.config.sidekiq_redis_url.to_s}
-  config.default_job_options = {"retry" => Miteru.config.sidekiq_job_retry}
+  config.default_job_options = {
+    retry: Miteru.config.sidekiq_job_retry,
+    expires_in: 0.second
+  }
 end
 
 Sidekiq.configure_client do |config|

lib/miteru/sidekiq/jobs.rb

@@ -1,6 +1,7 @@
 # frozen_string_literal: true
 
 require "sidekiq"
+require "timeout"
 
 module Miteru
   module Jobs
@@ -14,12 +15,15 @@ class CrawleJob
       #
       def perform(url, source)
         website = Miteru::Website.new(url, source:)
+
         with_db_connection do
-          result = Crawler.result(website)
-          if result.success?
-            Miteru.logger.info("Crawler:#{website.truncated_url} succeeded.")
-          else
-            Miteru.logger.info("Crawler:#{website.truncated_url} failed - #{result.failure}.")
+          Timeout.timeout(Miteru.config.sidekiq_job_timeout) do
+            result = Crawler.result(website)
+            if result.success?
+              Miteru.logger.info("Crawler:#{website.truncated_url} succeeded.")
+            else
+              Miteru.logger.info("Crawler:#{website.truncated_url} failed - #{result.failure}.")
+            end
           end
         end
       end