Restrict cache on forked prs (#480)

## 📝 Description

renderedtext/project-tasks#2647

## ✅ Checklist
- [x] I have tested this change
- [ ] This change requires documentation update: To be done

Author: VeljkoMaksimovic

zebra/lib/zebra/workers/job_request_factory.ex

@@ -63,17 +63,17 @@ defmodule Zebra.Workers.JobRequestFactory do
          {:ok, rsa} <- Task.await(gen_rsa),
          {:ok, project} <- Task.await(find_project),
          find_repository <- Task.async(fn -> Repository.find(project.repository_id) end),
-         find_cache <-
-           Task.async(fn ->
-             if Job.self_hosted?(job.machine_type),
-               do: {:ok, nil},
-               else: Cache.find(project.cache_id)
-           end),
          find_artifact_token <-
            Task.async(fn ->
              Artifacthub.generate_token(project.artifact_store_id, job.id, project.id, spec)
            end),
          {:ok, repo_proxy} <- Task.await(find_repo_proxy),
+         find_cache <-
+           Task.async(fn ->
+             if Job.self_hosted?(job.machine_type),
+               do: {:ok, nil},
+               else: Cache.find(project.cache_id, repo_proxy)
+           end),
          find_secrets <-
            Task.async(fn -> Secrets.load(org_id, job.id, spec, project, repo_proxy) end),
          {:ok, repository, private_git_key} <- Task.await(find_repository),

zebra/lib/zebra/workers/job_request_factory/cache.ex

@@ -13,19 +13,20 @@ defmodule Zebra.Workers.JobRequestFactory.Cache do
   # Overall, if cache system is down, we ignore every issue.
   #
 
-  def find(nil) do
+  def find(nil, _repo_proxy) do
     # We don't fail any jobs due to a missing cache connection,
     # but we should make sure we are aware of any issues in this area.
     Watchman.increment("external.cachehub.describe.failed")
 
     {:ok, nil}
   end
 
-  def find(cache_id) do
+  def find(cache_id, repo_proxy) do
     Watchman.benchmark("external.cachehub.describe", fn ->
       req = Request.new(cache_id: cache_id)
 
-      with {:ok, endpoint} <- Application.fetch_env(:zebra, :cachehub_api_endpoint),
+      with false <- forked_pr?(repo_proxy),
+           {:ok, endpoint} <- Application.fetch_env(:zebra, :cachehub_api_endpoint),
            {:ok, channel} <- GRPC.Stub.connect(endpoint),
            {:ok, response} <- Stub.describe(channel, req, timeout: 30_000) do
         if response.status.code == InternalApi.ResponseStatus.Code.value(:OK) do
@@ -38,6 +39,10 @@ defmodule Zebra.Workers.JobRequestFactory.Cache do
           {:ok, nil}
         end
       else
+        true ->
+          Logger.info("Skipping fetching of the cache as the job is part of Forked PR build.")
+          {:ok, nil}
+
         e ->
           Watchman.increment("external.cachehub.describe.failed")
           Logger.info("Failed to fetch info from cachehub #{cache_id}, #{inspect(e)}")
@@ -82,4 +87,12 @@ defmodule Zebra.Workers.JobRequestFactory.Cache do
       {:ok, vars}
     end
   end
+
+  defp forked_pr?(_repo = %{pr_slug: ""}), do: false
+
+  defp forked_pr?(repo) do
+    [pr_repo | _rest] = repo.pr_slug |> String.split("/")
+    [base_repo | _rest] = repo.repo_slug |> String.split("/")
+    pr_repo != base_repo
+  end
 end