Backend - Marking auto-added artifacts as optional

[Ark-kun]

This PR is implementing part of the #1250 plan.

Problem:
DSL-compiler has been auto-adding two artifacts to all pipeline tasks (including ones that do not actually outputs those artifacts). All out existing compiled pipelines have this problem.
Argo 2.3 makes missing output artifacts a reason for task failure.
The workaround for the existing pipelines is for backend to mark those two artifacts as optional.

P.S.
The optional property is only available starting with Argo 2.3, so I'm bumping the Argo version.

---

This change is 

[Ark-kun]

The tests are currently failing since they use the wrong Argo version (despite updated go.mod). @neuromage What do I need to change to make Bazel use the proper Argo version?

[neuromage]

You can run

bazel run //:gazelle -- --update-repos --from_file=go.mod

[Ark-kun]

You can run

bazel run //:gazelle -- --update-repos --from_file=go.mod

Thanks.
I did that yesterday, but bazel fails for me:

~/go/src/github.com/kubeflow/pipelines$ bazel run //:gazelle -- --update-repos --from_file=go.mod
Starting local Bazel server and connecting to it...
ERROR: /usr/local/google/home/avolkov/.cache/bazel/_bazel_avolkov/31e08d5c4be525ad16464e998010eee7/external/org_tensorflow/third_party/gpus/cuda_configure.bzl:115:1: load() statements must be called before any other statement. Firs
t non-load() statement appears at /usr/local/google/home/avolkov/.cache/bazel/_bazel_avolkov/31e08d5c4be525ad16464e998010eee7/external/org_tensorflow/third_party/gpus/cuda_configure.bzl:26:1. Use --incompatible_bzl_disallow_load_af
ter_statement=false to temporarily disable this check.
ERROR: error loading package '': in /usr/local/google/home/avolkov/.cache/bazel/_bazel_avolkov/31e08d5c4be525ad16464e998010eee7/external/org_tensorflow/tensorflow/workspace.bzl: Extension 'third_party/gpus/cuda_configure.bzl' has e
rrors
ERROR: error loading package '': in /usr/local/google/home/avolkov/.cache/bazel/_bazel_avolkov/31e08d5c4be525ad16464e998010eee7/external/org_tensorflow/tensorflow/workspace.bzl: Extension 'third_party/gpus/cuda_configure.bzl' has e
rrors
INFO: Elapsed time: 2.586s
INFO: 0 processes.
FAILED: Build did NOT complete successfully (0 packages loaded)
FAILED: Build did NOT complete successfully (0 packages loaded)

[Ark-kun]

TF has fixed the issue ~1 month ago tensorflow/tensorflow@494d2cc It's part of the v1.12.2 release, but not v1.13.1.

[Ark-kun]

All the tests are passing.

[neuromage]

/lgtm

[IronPan]

/lgtm

[IronPan]

/lgtm
/approve

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: IronPan

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [IronPan]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: IronPan

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [IronPan]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[Ark-kun]

@IronPan Looks like there is a problem in how Kubeflow deploys Argo. They deploy two workflow controllers with different configurations (one in argo namespace and another one in kubeflow namespaces). KF Pipelines seems to be using the one from argo namespace which is not configured using configMap.