Use errors middleware to redirect on idled workspace after 4 seconds

[dkwon17]

Signed-off-by: David Kwon [email protected]

What does this PR do?

Alternative PR to #1386. The main difference is that this PR uses the Errors Traefik middleware.

Screenshot/screencast of this PR

In this gif we see that:

1. Workspace is stopped
2. IDE page is refreshed
3. Instead of a 504 error, the user is redirected to the dashboard after 4 seconds

What issues does this PR fix or reference?

Part of eclipse-che/che#20599. In another PR, maybe we can redirect the user to some error page provided by the dashboard.

How to test this PR?

Tested on CRC with the following:

1. Deploy

chectl server:deploy \
     --platform crc\
     --che-operator-image quay.io/dkwon17/che-operator:useerrorsmiddleware

2. Open the dashboard with the URL provided by the output from chectl.
3. Start a workspace, and wait until the editor opens.
4. Stop the workspace.
5. Refresh the editor. You should be redirected to the dashboard after 4 seconds where you can start the workspace again.

PR Checklist

As the author of this Pull Request I made sure that:

• The Eclipse Contributor Agreement is valid
• Code produced is complete
• Code builds without errors
• Tests are covering the bugfix
• Custom resource definition file is up to date
• OLM bundles are up to date
• The repository devfile is up to date and works
• Sections What issues does this PR fix or reference and How to test this PR completed
• Relevant user documentation updated
• Relevant contributing documentation updated
• CI/CD changes implemented, documented and communicated

Reviewers

Reviewers, please comment how you tested the PR when approving it.

[openshift-ci]

Skipping CI for Draft Pull Request.
If you want CI signal for your change, please convert it to an actual PR.
You can still manually trigger a test run with /test all

[codecov]

Codecov Report

Merging #1392 (c130e99) into main (705d653) will increase coverage by 0.05%.
The diff coverage is 100.00%.

❗ Current head c130e99 differs from pull request most recent head 67182b0. Consider uploading reports for the commit 67182b0 to get more accurate results

@@            Coverage Diff             @@
##             main    #1392      +/-   ##
==========================================
+ Coverage   63.12%   63.18%   +0.05%     
==========================================
  Files          70       70              
  Lines        5804     5834      +30     
==========================================
+ Hits         3664     3686      +22     
- Misses       1777     1785       +8     
  Partials      363      363              

Impacted Files	Coverage Δ
controllers/devworkspace/solver/che_routing.go	79.52% <100.00%> (+0.55%)	⬆️
pkg/deploy/gateway/traefik_config_util.go	91.80% <100.00%> (+2.00%)	⬆️
pkg/common/operator-defaults/defaults.go	15.97% <0.00%> (-0.80%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 705d653...67182b0. Read the comment docs.

[benoitf]

I was wondering if we should redirect to the workspace page in dashboard instead of the home page.

[dkwon17]

The workspace page is <CHE_HOST>/dashboard/#/workspaces is that right?

On the Traeffik side with the Traeffik errors middleware, the content after the # is considered a URL fragment and is ignored because it is a browser-concept: traefik/traefik#9022 (comment)

To support redirection to the workspaces page, I think we would need to make some changes where the dashboard server receives the HTTP request.

[dkwon17]

There is a bug somewhere in this implementation. When restarting an idled workspace, instead of opening the IDE, I'm redirected to the dashboard. I am investigating

EDIT: the latest commit (3f85fd3) fixes this issue

[dkwon17]

This PR is ready to review. This PR adds two middlewares (errors and headers) to the workspace Traefik config:

Here is an example config diff reflecting what this PR adds:

		  middlewares:
		    workspace27366e668ca645c9-auth:
		      forwardAuth:
		        address: http://127.0.0.1:8089?namespace=admin-che
		        trustForwardHeader: false
+		    workspace27366e668ca645c9-errors:
+		      errors:
+		        query: /
+		        service: che-dashboard
+		        status: 500-599
+		    workspace27366e668ca645c9-headers:
+		      headers:
+		        customResponseHeaders:
+		          cache-control: no-store, max-age=0
		    workspace27366e668ca645c9-strip-prefix:
		      stripPrefix:
		        prefixes:
		        - /workspace27366e668ca645c9
		  routers:
		    workspace27366e668ca645c9:
		      middlewares:
		      - workspace27366e668ca645c9-strip-prefix
		      - workspace27366e668ca645c9-auth
+		      - workspace27366e668ca645c9-headers
+		      - workspace27366e668ca645c9-errors
		      priority: 100
		      rule: PathPrefix(`/workspace27366e668ca645c9`)
		      service: workspace27366e668ca645c9
		    workspace27366e668ca645c9-3100-healthz:
		      middlewares:
		      - workspace27366e668ca645c9-strip-prefix
		      priority: 101
		      rule: Path(`/workspace27366e668ca645c9/theia-ide/3100/healthz`)
		      service: workspace27366e668ca645c9
+		  serversTransports:
+		    workspace27366e668ca645c9:
+		      forwardingTimeouts:
+		        dialTimeout: 4s
		  services:
		    workspace27366e668ca645c9:
		      loadBalancer:
		        servers:
		        - url: http://workspace27366e668ca645c9-service.admin-che.svc:3030
		        serversTransport: workspace27366e668ca645c9

The headers middleware was added to deal with this issue with browser cache: #1392 (comment).

The serverTransport config was added to add a 4 second timeout for when trying to access the workspace url. If the workspace server cannot be reached after 4 seconds, then, the errors middleware will kick in and redirect requests to the workspace url to the dashboard.

[openshift-ci]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by: dkwon17, ibuziuk, tolusha

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

Needs approval from an approver in each of these files:

• OWNERS

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

[openshift-ci]

New changes are detected. LGTM label has been removed.

[dkwon17]

/retest

[openshift-ci]

@dkwon17: The following test failed, say /retest to rerun all failed tests or /retest-required to rerun all mandatory failed tests:

Test name	Commit	Details	Required	Rerun command
ci/prow/v9-devworkspace-happy-path	ef96f57	link	true	/test v9-devworkspace-happy-path

Full PR test history. Your PR dashboard.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. I understand the commands that are listed here.

[dkwon17]

/retest

[tolusha]

@dkwon17
v9-devworkspace-happy-path is unstable.
I checked and results look good.
Pls merge.

[dkwon17]

Thank you @tolusha , will merge