Pass commit data to ProcessCommitWorkerprocess-commit-worker-improvements

By passing commit data to this worker we remove the need for querying
the Git repository for every job. This in turn reduces the time spent
processing each job.

The migration included migrates jobs from the old format to the new
format. For this to work properly it requires downtime as otherwise
workers may start producing errors until they're using a newer version
of the worker code.

3 files changed, 21 insertions, 10 deletions

diff --git a/app/models/commit.rb b/app/models/commit.rb
index 9e7fde9503d..176c524cf7b 100644
--- a/app/models/commit.rb
+++ b/app/models/commit.rb
@@ -48,6 +48,10 @@ class Commit
         max_lines: DIFF_HARD_LIMIT_LINES,
       }
     end
+
+    def from_hash(hash, project)
+      new(Gitlab::Git::Commit.new(hash), project)
+    end
   end
 
   attr_accessor :raw
diff --git a/app/services/git_push_service.rb b/app/services/git_push_service.rb
index 647930d555c..185556c12cc 100644
--- a/app/services/git_push_service.rb
+++ b/app/services/git_push_service.rb
@@ -135,7 +135,7 @@ class GitPushService < BaseService
 
     @push_commits.each do |commit|
       ProcessCommitWorker.
-        perform_async(project.id, current_user.id, commit.id, default)
+        perform_async(project.id, current_user.id, commit.to_hash, default)
     end
   end
 
diff --git a/app/workers/process_commit_worker.rb b/app/workers/process_commit_worker.rb
index 071741fbacd..e9a5bd7f24e 100644
--- a/app/workers/process_commit_worker.rb
+++ b/app/workers/process_commit_worker.rb
@@ -10,9 +10,10 @@ class ProcessCommitWorker
 
   # project_id - The ID of the project this commit belongs to.
   # user_id - The ID of the user that pushed the commit.
-  # commit_sha - The SHA1 of the commit to process.
+  # commit_hash - Hash containing commit details to use for constructing a
+  #               Commit object without having to use the Git repository.
   # default - The data was pushed to the default branch.
-  def perform(project_id, user_id, commit_sha, default = false)
+  def perform(project_id, user_id, commit_hash, default = false)
     project = Project.find_by(id: project_id)
 
     return unless project
@@ -21,10 +22,7 @@ class ProcessCommitWorker
 
     return unless user
 
-    commit = find_commit(project, commit_sha)
-
-    return unless commit
-
+    commit = build_commit(project, commit_hash)
     author = commit.author || user
 
     process_commit_message(project, commit, user, author, default)
@@ -59,9 +57,18 @@ class ProcessCommitWorker
       update_all(first_mentioned_in_commit_at: commit.committed_date)
   end
 
-  private
+  def build_commit(project, hash)
+    date_suffix = '_date'
+
+    # When processing Sidekiq payloads various timestamps are stored as Strings.
+    # Commit in turn expects Time-like instances upon input, so we have to
+    # manually parse these values.
+    hash.each do |key, value|
+      if key.to_s.end_with?(date_suffix) && value.is_a?(String)
+        hash[key] = Time.parse(value)
+      end
+    end
 
-  def find_commit(project, sha)
-    project.commit(sha)
+    Commit.from_hash(hash, project)
   end
 end