perf: remove no-remote-cache execution requirements (#2043)

These were introduced to reduce load on a remote-cache instance to avoid network saturation.
A month later, a feature was added in one remote-cache implementatation which provides a different fix:
buchgr/bazel-remote#440 rejects large input files on upload.

In practice, while these action do often produce huge outputs, they are also slow to re-execute.
In many cases it's worth it to use a remote-cache for RunAndCommitLayer in particular to avoid a local rebuild
even though it's a large network fetch.
Currently users can't configure this because we've hardcoded the values. If they do want to keep the
no-remote-cache execution requirement, they can do this via a tag (provided they opt-in to
experimental_allow_tags_propagation, see bazelbuild/bazel#8830)

#1856 (comment) is an example of a user
asking for these to be removed.

container/layer.bzl

@@ -166,11 +166,6 @@ def build_layer(
         input_manifests = xz_input_manifests,
         tools = files + file_map.values() + tars + debs + [manifest_file] + xz_tools,
         outputs = [layer],
-        execution_requirements = {
-            # This action produces large output files, so it's not
-            # economical to send this to the remote-cache
-            "no-remote-cache": "1",
-        },
         use_default_shell_env = True,
         mnemonic = "ImageLayer",
     )

container/layer_tools.bzl

@@ -187,12 +187,6 @@ def assemble(
         arguments = [args],
         tools = inputs,
         outputs = [output],
-        execution_requirements = {
-            # This action produces large output files, but doesn't require much CPU to compute.
-            # It's not economical to send this to the remote-cache, instead local cache misses
-            # should just run join_layers again.
-            "no-remote-cache": "1",
-        },
         mnemonic = "JoinLayers",
     )
 

docker/package_managers/install_pkgs.bzl

@@ -131,11 +131,6 @@ def _impl(ctx, image_tar = None, installables_tar = None, installation_cleanup_c
             installables_tar,
             image_util,
         ],
-        execution_requirements = {
-            # This action produces large output files, and isn't economical to
-            # upload to a remote cache.
-            "no-remote-cache": "1",
-        },
         mnemonic = "ExtractImageId",
         tools = [ctx.executable._extract_image_id, ctx.executable._to_json_tool],
         executable = script,

docker/util/run.bzl

@@ -396,11 +396,6 @@ def _commit_layer_impl(
         outputs = [output_layer_tar, output_diff_id],
         inputs = runfiles,
         executable = script,
-        execution_requirements = {
-            # This action produces large output files, and isn't economical to
-            # upload to a remote cache.
-            "no-remote-cache": "1",
-        },
         mnemonic = "RunAndCommitLayer",
         tools = [ctx.executable._extract_image_id, ctx.executable._last_layer_extractor_tool],
         use_default_shell_env = True,

skylib/zip.bzl

@@ -50,12 +50,6 @@ def _gzip(ctx, artifact, out, decompress, options, mnemonic):
             ) + (options or []),
             inputs = [artifact],
             outputs = [out],
-            execution_requirements = {
-                # This action produces large output files, but doesn't require much CPU to compute.
-                # It's not economical to send this to the remote-cache, instead local cache misses
-                # should just run gzip again.
-                "no-remote-cache": "1",
-            },
             mnemonic = mnemonic,
             tools = ctx.attr._zipper[DefaultInfo].default_runfiles.files,
         )
@@ -75,13 +69,6 @@ def _gzip(ctx, artifact, out, decompress, options, mnemonic):
             input_manifests = input_manifests,
             inputs = [artifact],
             outputs = [out],
-            execution_requirements = {
-                # See comment above.
-                # In this case it's not guaranteed that the user's gzip implementation
-                # is actually faster than a network fetch, but we guess that any typical
-                # one should be.
-                "no-remote-cache": "1",
-            },
             use_default_shell_env = True,
             mnemonic = mnemonic,
             tools = tools,