[codex] Structure managed endpoint allocation failures

[juliusmarminge]

Summary

• attach operation, user, environment, resource identifiers, and exact causes to managed-endpoint allocation, provisioning, deprovisioning, tunnel-client, and DNS-client failures
• replace the synthetic reserve Error and valueless persistence-error constructor wrapper with structural errors; invalid tunnel responses are a distinct cause-free error nested under the provisioning failure
• use Effect.catchTags for tunnel/DNS recovery and only treat structured nested NotFound failures as idempotent fallback
• preserve unrelated checkpoint update failures instead of silently retrying through a different DNS path
• export the real service make values

Verification

• vp test infra/relay/src/environments/ManagedEndpointAllocations.test.ts infra/relay/src/environments/ManagedEndpointProvider.test.ts (19 passed)
• vp check (passes with 20 unrelated existing warnings)
• vp run typecheck

Scope

Excludes relay orchestration/projection and files owned by #3331, #3334, #3335, and #3392.

---

Note

Medium Risk
Changes live in Cloudflare tunnel/DNS provisioning paths; misclassified errors could break retries or hide outages, though behavior is narrowed and tested for non-not-found checkpoint failures.

Overview
Tightens managed endpoint provisioning so transient DNS “not found” cases still fall back, but real failures surface correctly.

Checkpoint DNS updates no longer swallow every updateRecord error via orElseSucceed; only structured NotFound causes are treated as “record gone, reconcile elsewhere.” Other errors (e.g. Cloudflare outage) fail at ensure-dns-record instead of silently taking another DNS path—covered by a new test.

Tunnel validation now requires the created tunnel’s name to match the expected stable name (not merely non-null), with returnedTunnelName on the provisioning failure when it mismatches.

Exports and layers: make is exported on allocations and provider; tunnel/DNS layer* helpers pass client implementations directly without extra of() wrappers; Cloudflare binding layer is flattened accordingly. Test fakes use tagged NotFound causes for missing DNS records.

^{Reviewed by Cursor Bugbot for commit 10793eb. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Harden managed endpoint allocation failure handling in ManagedEndpointProvider

• ensureDnsRecord now rethrows DNS updateRecord errors unless the cause is NotFound; previously all errors were swallowed and treated as a miss, masking real failures
• Tunnel response validation now checks that the returned tunnel name matches the expected name, not just that it is truthy; mismatches fail with stage validate-tunnel-response and include returnedTunnelName in the error
• Exports the make factory from both ManagedEndpointProvider and ManagedEndpointAllocations for external construction
• Removes the makeTunnelClient/makeDnsClient wrapper functions; service objects are now passed directly to layerTunnelClient/layerDnsClient
• Risk: the ensureDnsRecord change is a behavioral breaking change — callers that previously succeeded despite DNS update errors will now see failures propagated

^{Macroscope summarized 10793eb.}

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro

Run ID: b2b8b8af-6487-47ab-8b30-7559165ad79e

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch codex/relay-environments-error-audit

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[macroscopeapp]

Approvability

Verdict: Approved

Bug fix that improves error handling by properly surfacing DNS client failures instead of silently swallowing them. Changes include stricter tunnel response validation and mechanical removal of unnecessary wrapper functions, all covered by tests.

^{You can customize Macroscope's approvability policy. Learn more.}

Dismissing prior approval to re-evaluate 300483c

Dismissing prior approval to re-evaluate 10793eb