Add workspace-lifecycle real-infra staginge2e (core#2332 P1.10) #2338

Merged

claude-ceo-assistant merged 1 commits from core2332-p110-workspace-lifecycle-staginge2e into main 2026-06-06 05:20:55 +00:00

Conversation 0 Commits 1 Files Changed 3

+752

hongming-codex-laptop commented 2026-06-06 04:47:13 +00:00

Member

Copy Link

Copy Source

What

Close the workspace-lifecycle coverage gap (core#2332 P1.10): soft-restart / pause / resume / hibernate were only unit-tested (httptest in workspace-server/internal/handlers/*_test.go) and never proven against a real container.

New Go suite workspace-server/internal/staginge2e (//go:build staging_e2e), mirroring the cp internal/staginge2e harness idioms (cp#386): STAGING_E2E=1 gate, CP_ADMIN_API_TOKEN admin surface, provision → wait-online → assert, t.Cleanup teardown. molecule-core has no CP client packages, so the harness is HTTP-only and self-contained (no go.moleculesai.app/controlplane/... imports).

What each transition asserts (observable state, not just HTTP 200)

TestWorkspaceLifecycle_Staging provisions a real throwaway staging tenant + workspace, then:

Endpoint	Asserts
restart POST /restart	body {"status":"provisioning"}; then live GET /workspaces/:id → online + routable url; post-restart A2A serve probe returns 2xx (container actually came back, not just a row flip)
pause POST /pause	→ paused; url cleared; A2A no longer serves — the genuinely-stopped signal (a flag-only handler would still be reachable)
resume POST /resume	paused → provisioning → online + routable; serveable again
hibernate POST /hibernate?force=true	→ hibernated (settled, not stuck mid-hibernating); url cleared; unserveable
wake	next A2A message auto-wakes hibernated → online (Resume only handles paused); serveable again

Status is read from the live DB-backed GET /workspaces/:id — the lifecycle POST body could lie; the GET proves the row. The restart provisioning window is observed non-fatally (a fast box can race back to online before the first poll); the load-bearing assertions are eventual online+routable + a successful serve probe.

Honest limit (precise TODO left in code)

The strongest "container stopped" signal is the EC2/Docker power-state, which is only observable CP-side (AWS/SSM) and not reachable from the core ws-server module without importing the CP client surface. assertNotServing asserts the strongest signal available here (url cleared + immediate non-serve) with a TODO(core#2332) to tighten to instance power-state if a CP admin endpoint ever surfaces it to the tenant API. Nothing was weakened to make this pass.

CI

New workflow e2e-workspace-lifecycle.yml:

• PR path = cheap, honest compile+skip gate: go vet -tags staging_e2e + assert the suite SKIPs LOUD without creds. NOT a fake-green mask (a broken test file fails at PR time). Non-required.
• Real run = workflow_dispatch / schedule (daily 08:00 UTC, offset from e2e-staging-saas 07:00 + peer-visibility 07:30) with CP_BASE_URL + CP_STAGING_ADMIN_API_TOKEN. Teardown via test t.Cleanup; the age-guarded sweep-stale-e2e-orgs (30-min floor, e2e- prefix) is the final net.

Advisory-by-infra: this suite needs a live staging tenant, so it is intentionally NOT a merge-blocking required check. Promote-to-required is a separate CTO decision (mirrors cp#386 and the peer-visibility flip-to-required pattern, molecule-core#1296).

Validation

go vet -tags staging_e2e ./internal/staginge2e/...        # clean
go test -tags staging_e2e ./internal/staginge2e/ \
  -run TestWorkspaceLifecycle -count=1                     # compiles + SKIPs loud (no creds)
gofmt -l ...                                               # clean
go test ./...                                              # package excluded (tag-gated)

No self-merge — routed to agent-reviewer-cr2 + agent-researcher.

🤖 Generated with Claude Code

## What Close the workspace-lifecycle coverage gap (**core#2332 P1.10**): soft-restart / pause / resume / hibernate were only unit-tested (httptest in `workspace-server/internal/handlers/*_test.go`) and never proven against a real container. New Go suite **`workspace-server/internal/staginge2e`** (`//go:build staging_e2e`), mirroring the cp `internal/staginge2e` harness idioms (cp#386): `STAGING_E2E=1` gate, `CP_ADMIN_API_TOKEN` admin surface, provision → wait-online → assert, `t.Cleanup` teardown. molecule-core has no CP client packages, so the harness is **HTTP-only and self-contained** (no `go.moleculesai.app/controlplane/...` imports). ## What each transition asserts (observable state, not just HTTP 200) `TestWorkspaceLifecycle_Staging` provisions a real throwaway staging tenant + workspace, then: | Endpoint | Asserts | |---|---| | **restart** `POST /restart` | body `{"status":"provisioning"}`; then live `GET /workspaces/:id` → `online` + routable `url`; **post-restart A2A serve probe returns 2xx** (container actually came back, not just a row flip) | | **pause** `POST /pause` | `→ paused`; **`url` cleared**; **A2A no longer serves** — the genuinely-stopped signal (a flag-only handler would still be reachable) | | **resume** `POST /resume` | `paused → provisioning → online` + routable; **serveable again** | | **hibernate** `POST /hibernate?force=true` | `→ hibernated` (settled, not stuck mid-`hibernating`); `url` cleared; unserveable | | **wake** | next A2A message auto-wakes `hibernated → online` (Resume only handles `paused`); serveable again | Status is read from the live DB-backed `GET /workspaces/:id` — the lifecycle POST body could lie; the GET proves the row. The restart provisioning window is observed **non-fatally** (a fast box can race back to online before the first poll); the load-bearing assertions are eventual online+routable + a successful serve probe. ### Honest limit (precise TODO left in code) The strongest "container stopped" signal is the EC2/Docker power-state, which is only observable **CP-side** (AWS/SSM) and not reachable from the core ws-server module without importing the CP client surface. `assertNotServing` asserts the strongest signal available here (`url` cleared + immediate non-serve) with a `TODO(core#2332)` to tighten to instance power-state if a CP admin endpoint ever surfaces it to the tenant API. **Nothing was weakened to make this pass.** ## CI New workflow `e2e-workspace-lifecycle.yml`: - **PR path** = cheap, honest **compile+skip** gate: `go vet -tags staging_e2e` + assert the suite SKIPs LOUD without creds. NOT a fake-green mask (a broken test file fails at PR time). **Non-required.** - **Real run** = `workflow_dispatch` / `schedule` (daily 08:00 UTC, offset from e2e-staging-saas 07:00 + peer-visibility 07:30) with `CP_BASE_URL` + `CP_STAGING_ADMIN_API_TOKEN`. Teardown via test `t.Cleanup`; the age-guarded `sweep-stale-e2e-orgs` (30-min floor, `e2e-` prefix) is the final net. **Advisory-by-infra:** this suite needs a live staging tenant, so it is intentionally NOT a merge-blocking required check. **Promote-to-required is a separate CTO decision** (mirrors cp#386 and the peer-visibility flip-to-required pattern, molecule-core#1296). ## Validation ``` go vet -tags staging_e2e ./internal/staginge2e/... # clean go test -tags staging_e2e ./internal/staginge2e/ \ -run TestWorkspaceLifecycle -count=1 # compiles + SKIPs loud (no creds) gofmt -l ... # clean go test ./... # package excluded (tag-gated) ``` No self-merge — routed to agent-reviewer-cr2 + agent-researcher. 🤖 Generated with [Claude Code](https://claude.com/claude-code)

hongming-codex-laptop added 1 commit 2026-06-06 04:47:16 +00:00

Add workspace-lifecycle real-infra staginge2e (core#2332 P1.10) ... 7e313d1c77

Close the workspace-lifecycle coverage gap: soft-restart / pause / resume
/ hibernate were only unit-tested (httptest in
workspace-server/internal/handlers/*_test.go) and never proven against a
real container.

New Go suite workspace-server/internal/staginge2e (build tag
//go:build staging_e2e), mirroring the cp internal/staginge2e idioms
(cp#386): STAGING_E2E=1 gate, CP_ADMIN_API_TOKEN admin surface,
provision -> wait-online -> assert, t.Cleanup teardown. Core has no CP
client packages, so the harness is HTTP-only and self-contained.

TestWorkspaceLifecycle_Staging provisions a real throwaway staging tenant
+ workspace, then drives each lifecycle endpoint and asserts OBSERVABLE
state (not just HTTP 200):

- restart  -> body provisioning, then GET status -> online+routable, and a
             post-restart A2A serve probe succeeds (container actually back).
- pause    -> status paused + url cleared + workspace no longer serves A2A
             (the genuinely-stopped signal: a flag-only handler would still
             serve). resume -> online + serveable again.
- hibernate-> status hibernated + url cleared + unserveable; wake via the
             next A2A message -> online + serveable (auto-wake-on-message;
             Resume only handles paused).

Status is read from the live DB-backed GET /workspaces/:id (the lifecycle
POST body could lie; the GET proves the row). The restart provisioning
window is observed non-fatally (a fast box can race back to online before
the first poll) — the load-bearing assertions are eventual online+routable
and a successful serve probe.

The strongest "container stopped" signal is EC2/Docker power-state, only
observable CP-side (AWS/SSM) and not reachable from the core ws-server
module; assertNotServing asserts the strongest signal available here
(url cleared + immediate non-serve) with a precise TODO(core#2332).

Advisory-by-infra: the real run needs a live staging tenant, so the new
workflow e2e-workspace-lifecycle.yml runs it on workflow_dispatch /
schedule only (daily 08:00 UTC, offset from the other staging e2es). The
PR path is a cheap honest compile+skip gate (vet under the tag + assert it
SKIPs LOUD without creds) — NOT required. Promote-to-required is a
separate CTO decision (mirrors cp#386 / the peer-visibility flip pattern,
molecule-core#1296).

Validation: go vet -tags staging_e2e ./internal/staginge2e/... (clean);
go test -tags staging_e2e ./internal/staginge2e/ -run TestWorkspaceLifecycle
-count=1 compiles and SKIPs loud without creds; gofmt clean; default
`go test ./...` excludes the package (tag-gated).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

hongming-codex-laptop requested review from agent-reviewer-cr2 2026-06-06 04:47:53 +00:00

hongming-codex-laptop requested review from agent-researcher 2026-06-06 04:47:54 +00:00

claude-ceo-assistant merged commit a85d4c8f89 into main 2026-06-06 05:20:55 +00:00

Sign in to join this conversation.

Reviewers

No Reviewers

agent-reviewer-cr2

agent-researcher

Labels

Clear labels

area/ci

CI/CD pipeline issues

do-not-auto-merge

Opt out of autonomous merge-queue merging

kind/infrastructure

Infrastructure-related issues

merge-queue

Ready for serialized Gitea merge queue

merge-queue-hold

Temporarily hold PR in merge queue

platform/go

Go platform test issues

release-blocker

Blocks the staging→main promotion / a release

release-test

security

test-label-sre

tier:high

High risk per dev-sop §SOP-6 — ceo only, 24h cooldown

tier:low

Low risk per dev-sop §SOP-6 — engineers/managers/ceo can approve

tier:medium

Medium risk per dev-sop §SOP-6 — managers/ceo can approve

triage-test

test

wip

Work in progress — do not auto-merge

No Label

Milestone

No items

No Milestone

Projects

Clear projects

No project

Assignees

Clear assignees

agent-dev-a agent-dev-b agent-pm agent-researcher agent-reviewer agent-reviewer-1 agent-reviewer-cr2 app-fe (Molecule AI · app-fe) app-lead (Molecule AI · app-lead) app-qa (Molecule AI · app-qa) claude-ceo-assistant claude-ci-reader claude-status-reaper core-be (Molecule AI · core-be) core-devops (Molecule AI · core-devops) core-fe (Molecule AI · core-fe) core-lead (Molecule AI · core-lead) core-offsec (Molecule AI · core-offsec) core-qa (Molecule AI · core-qa) core-security (Molecule AI · core-security) core-uiux (Molecule AI · core-uiux) cp-be (Molecule AI · cp-be) cp-lead (Molecule AI · cp-lead) cp-qa (Molecule AI · cp-qa) cp-security (Molecule AI · cp-security) cui (Zhanlin Cui) dev-lead (Molecule AI · dev-lead) devops-engineer documentation-specialist (Molecule AI · documentation-specialist) fullstack-engineer (Molecule AI · fullstack-engineer) hongming hongming-ceo-delegated hongming-codex-laptop hongming-kimi-laptop hongming-pc2 infra-lead (Molecule AI · infra-lead) infra-runtime-be (Molecule AI · infra-runtime-be) infra-sre (Molecule AI · infra-sre) integration-tester (Molecule AI · integration-tester) mc-drift-bot molecule-code-reviewer plugin-dev (Molecule AI · plugin-dev) pm publish-runtime-bot pypi-publisher (Molecule AI PyPI Publisher (RFC#596)) release-manager (Molecule AI · release-manager) sdk-dev (Molecule AI · sdk-dev) sdk-lead (Molecule AI · sdk-lead) sop-drift-bot sop-tier-bot (SOP Tier-Check Bot) technical-writer (Molecule AI · technical-writer) triage-operator (Molecule AI · triage-operator)

No Assignees

1 Participants

Notifications

Subscribe

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: molecule-ai/molecule-core#2338