fix: validate messaging compatible endpoint routing#2849
Merged
Conversation
Contributor
|
Note Reviews pausedIt looks like this branch is under active development. To avoid overwhelming you with review comments due to an influx of new commits, CodeRabbit has automatically paused this review. You can configure this behavior by changing the Use the following commands to manage reviews:
Use the checkboxes below for quick actions:
📝 WalkthroughWalkthroughAdds a messaging-compatible-endpoint E2E job and hermetic test script, onboarding smoke checks that validate managed inference routing and /chat/completions, a Telegram diagnostics Node preload injected at startup, test coverage additions, and PR template removals for AI disclosure. (≤50 words) Changes
Sequence Diagram(s)sequenceDiagram
participant User
participant Onboard as Onboarding Flow
participant Gateway
participant Sandbox as NEMOCLAW Sandbox
participant Inference as inference.local
participant Mock as OpenAI Mock Server
User->>Onboard: Start onboarding (compatible-endpoint)
Onboard->>Gateway: Query provider presence
Onboard->>Sandbox: Generate & run smoke test script
Sandbox->>Sandbox: Read /sandbox/.openclaw/openclaw.json
Sandbox->>Inference: POST /v1/chat/completions
Inference->>Mock: Forward authenticated request
Mock-->>Inference: Return mock response with marker
Inference-->>Sandbox: Return response
Sandbox->>Onboard: Report smoke PASS/FAIL
Onboard-->>User: Continue or abort onboarding
sequenceDiagram
participant E2E as E2E Runner
participant Mock as OpenAI Mock Server
participant Provisioner as NEMOCLAW Provisioner
participant Sandbox as NEMOCLAW Sandbox
participant Inference as inference.local
participant Telegram as Telegram API
E2E->>Mock: Boot mock server (Bearer auth)
E2E->>Provisioner: Provision sandbox with compatible-endpoint provider
Provisioner->>Sandbox: Initialize openclaw.json (managed provider -> inference.local)
E2E->>Sandbox: Check onboard logs for smoke marker
E2E->>Sandbox: Inspect openclaw.json provider shape
Sandbox->>Inference: Call /v1/chat/completions
Inference->>Mock: Authenticated POST logged
Mock-->>Inference: Return mock assistant content
Sandbox-->>E2E: Validate assistant content
E2E->>Telegram: (optional) perform Telegram round-trip
E2E-->>E2E: Aggregate PASS/FAIL/SKIP
Estimated code review effort🎯 4 (Complex) | ⏱️ ~45 minutes Poem
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Warning Review ran into problems🔥 ProblemsTimed out fetching pipeline failures after 30000ms Comment |
ericksoa
added
bug
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Actionable comments posted: 4
🧹 Nitpick comments (2)
test/nemoclaw-start.test.ts (1)
1129-1159: ⚡ Quick winExercise the Telegram preload behavior, not just its source text.
These assertions prove the heredoc is present, but they will not catch a broken sanitizer or a misleading startup-order breadcrumb. This file already executes extracted preloads for the safety net and Slack guard; adding one behavior-level Telegram test here would lock down the critical runtime behavior.
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@test/nemoclaw-start.test.ts` around lines 1129 - 1159, Add a behavior-level test that actually executes the Telegram diagnostics preload instead of only checking its text: locate the extracted heredoc for _TELEGRAM_DIAGNOSTICS_SCRIPT in the test src, write its contents to a temp file, ensure permissions via the same validate_tmp_permissions flow, then invoke the entrypoint path that calls install_telegram_diagnostics (the same path matched by configure_messaging_channels -> install_telegram_diagnostics) or simulate the connect-shell sourcing behavior so the temp script is required (NODE_OPTIONS/--require) and assert a side effect (env var, temp file touch, or stdout marker) the preload sets; this verifies the sanitizer and startup ordering actually run rather than just existing in the source.test/onboard.test.ts (1)
292-302: ⚡ Quick winAlso assert the smoke marker/prompt contract in the generated script
This test already checks routing and secret-safety well; adding assertions for the explicit probe prompt and marker would better lock the intended smoke behavior.
Suggested addition
it("builds a compatible-endpoint smoke script that validates managed inference config", () => { const script = buildCompatibleEndpointSandboxSmokeScript("deepseek-ai/DeepSeek-V4-Flash"); assert.match(script, /models\.providers\.inference/); assert.match(script, /https:\/\/inference\.local\/v1/); assert.match(script, /apiKey.*unused/); assert.match(script, /agents\.defaults\.model\.primary/); + assert.match(script, /Reply with exactly: PONG/); + assert.match(script, /INFERENCE_SMOKE_OK/); assert.match(script, /curl[\s\S]*\/chat\/completions/); assert.doesNotMatch(script, /COMPATIBLE_API_KEY/); assert.doesNotMatch(script, /api\.deepinfra\.com/); });🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@test/onboard.test.ts` around lines 292 - 302, The test for buildCompatibleEndpointSandboxSmokeScript should also assert the smoke probe prompt and marker contract are present in the generated script; update the test (in onboard.test.ts) that calls buildCompatibleEndpointSandboxSmokeScript("deepseek-ai/DeepSeek-V4-Flash") to add assertions that the script contains the explicit probe prompt text and the smoke marker/identifier used by the runtime (e.g., the probe prompt string and the SMOKE marker token your system expects), so add assert.match checks against those two unique patterns (the probe prompt and the smoke marker) to lock the intended smoke behavior.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@scripts/nemoclaw-start.sh`:
- Around line 866-872: The sanitize function's final regex only matches tokens
composed of a limited charset and stops at colons, leaking colon-delimited
tokens (e.g., Telegram tokens); update the replace call in sanitize (function
sanitize) to consume the entire token value up to whitespace or quotes by
replacing the right-hand character class with a greedy "not whitespace or quote"
pattern (e.g., use [^\s"']+ or similar) so patterns like
(api[_-]?key|token|authorization)["'=:\s]+[^\s"']+ are fully redacted; keep the
/gi flags for case-insensitivity.
- Around line 862-949: The stderr hook can emit a Telegram-specific failure
before Telegram is known ready; change the logic in process.stderr.write (the
handler that checks inferenceLogged and tests text for LLM errors) to only set
inferenceLogged and call emit(...) when readyLogged is true (i.e., gate the
breadcrumb on readyLogged). Locate process.stderr.write, describeRequest,
maybeLogTelegramReady and readyLogged/inferenceLogged, and add the check (skip
early if !readyLogged) so the “[telegram] [default] agent turn failed after
provider startup…” message is only emitted after maybeLogTelegramReady has set
readyLogged.
In `@src/lib/onboard.ts`:
- Around line 8541-8551: The smoke check is using session/remembered channels
(selectedMessagingChannels/recordedMessagingChannels) which can include channels
disabled by createSandbox; change the source to the sandbox's active channels by
retrieving registry.getSandbox(sandboxName)?.messagingChannels and pass that (or
an empty array if undefined) as messagingChannels into
verifyCompatibleEndpointSandboxSmoke instead of smokeMessagingChannels built
from selectedMessagingChannels/recordedMessagingChannels; keep the call to
verifyCompatibleEndpointSandboxSmoke and ensure
sandboxName/provider/model/endpointUrl/credentialEnv/agent are preserved.
In `@test/onboard.test.ts`:
- Around line 44-45: The runtime guard is missing checks for the two new helpers
on OnboardTestInternals; update isOnboardTestInternals to validate that
buildCompatibleEndpointSandboxSmokeScript and buildSandboxConfigSyncScript exist
and are functions (or match their expected types), so the guard fails early with
a clear error instead of causing "is not a function" at runtime when callers of
those helpers (e.g., where OnboardTestInternals is used) invoke them; locate the
isOnboardTestInternals function and add explicit type checks for these two
symbols to keep the guard in sync with the OnboardTestInternals interface.
---
Nitpick comments:
In `@test/nemoclaw-start.test.ts`:
- Around line 1129-1159: Add a behavior-level test that actually executes the
Telegram diagnostics preload instead of only checking its text: locate the
extracted heredoc for _TELEGRAM_DIAGNOSTICS_SCRIPT in the test src, write its
contents to a temp file, ensure permissions via the same
validate_tmp_permissions flow, then invoke the entrypoint path that calls
install_telegram_diagnostics (the same path matched by
configure_messaging_channels -> install_telegram_diagnostics) or simulate the
connect-shell sourcing behavior so the temp script is required
(NODE_OPTIONS/--require) and assert a side effect (env var, temp file touch, or
stdout marker) the preload sets; this verifies the sanitizer and startup
ordering actually run rather than just existing in the source.
In `@test/onboard.test.ts`:
- Around line 292-302: The test for buildCompatibleEndpointSandboxSmokeScript
should also assert the smoke probe prompt and marker contract are present in the
generated script; update the test (in onboard.test.ts) that calls
buildCompatibleEndpointSandboxSmokeScript("deepseek-ai/DeepSeek-V4-Flash") to
add assertions that the script contains the explicit probe prompt text and the
smoke marker/identifier used by the runtime (e.g., the probe prompt string and
the SMOKE marker token your system expects), so add assert.match checks against
those two unique patterns (the probe prompt and the smoke marker) to lock the
intended smoke behavior.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Enterprise
Run ID: c5b1e550-4cd0-4cbd-961a-3368d3021789
📒 Files selected for processing (10)
.coderabbit.yaml.github/workflows/nightly-e2e.yamlscripts/nemoclaw-start.shsrc/lib/inference-config.test.tssrc/lib/onboard.tstest/e2e/brev-e2e.test.tstest/e2e/test-messaging-compatible-endpoint.shtest/generate-openclaw-config.test.tstest/nemoclaw-start.test.tstest/onboard.test.ts
Contributor
Selective E2E Results — ✅ All requested jobs passedRun: 25224181432
|
Contributor
Selective E2E Results — ❌ Some jobs failedRun: 25224419429
|
Contributor
There was a problem hiding this comment.
♻️ Duplicate comments (1)
src/lib/onboard.ts (1)
8552-8561:⚠️ Potential issue | 🟠 Major | ⚡ Quick winUse active registry messaging channels for smoke gating
At Line 8552, smoke gating still falls back to selected/session channels. If a channel is remembered but disabled in the created/reused sandbox, this can incorrectly run the hard-fail smoke check and abort onboarding.
Suggested fix
- const smokeMessagingChannels = - selectedMessagingChannels.length > 0 ? selectedMessagingChannels : recordedMessagingChannels; + const activeRegistryChannels = registry.getSandbox(sandboxName)?.messagingChannels; + const smokeMessagingChannels = Array.isArray(activeRegistryChannels) + ? activeRegistryChannels + : selectedMessagingChannels.length > 0 + ? selectedMessagingChannels + : recordedMessagingChannels;🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@src/lib/onboard.ts` around lines 8552 - 8561, The smoke gating currently uses smokeMessagingChannels which falls back to selectedMessagingChannels or recordedMessagingChannels and can include channels that are disabled in the created/reused sandbox; change the call to verifyCompatibleEndpointSandboxSmoke to pass the sandbox's active messaging channels instead (derive an activeMessagingChannels list from the sandbox/registry state used when creating/reusing the sandbox, filtering out any channels marked disabled) so the smoke check uses only channels actually enabled in the sandbox (replace usage of smokeMessagingChannels with activeMessagingChannels and ensure any helper that reads recordedMessagingChannels/selectedMessagingChannels is updated to prefer sandbox/registry active channels).
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Duplicate comments:
In `@src/lib/onboard.ts`:
- Around line 8552-8561: The smoke gating currently uses smokeMessagingChannels
which falls back to selectedMessagingChannels or recordedMessagingChannels and
can include channels that are disabled in the created/reused sandbox; change the
call to verifyCompatibleEndpointSandboxSmoke to pass the sandbox's active
messaging channels instead (derive an activeMessagingChannels list from the
sandbox/registry state used when creating/reusing the sandbox, filtering out any
channels marked disabled) so the smoke check uses only channels actually enabled
in the sandbox (replace usage of smokeMessagingChannels with
activeMessagingChannels and ensure any helper that reads
recordedMessagingChannels/selectedMessagingChannels is updated to prefer
sandbox/registry active channels).
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Enterprise
Run ID: 85e66d73-035e-428d-84fc-be7808baa97f
📒 Files selected for processing (2)
src/lib/onboard.tstest/onboard.test.ts
Contributor
Selective E2E Results — ❌ Some jobs failedRun: 25225179236
|
Contributor
Selective E2E Results — ✅ All requested jobs passedRun: 25225962098
|
Contributor
Selective E2E Results — ✅ All requested jobs passedRun: 25226844478
|
Contributor
There was a problem hiding this comment.
♻️ Duplicate comments (2)
scripts/nemoclaw-start.sh (2)
948-956:⚠️ Potential issue | 🟠 Major | ⚡ Quick winOnly emit the inference breadcrumb after Telegram is actually ready.
Line 948 marks the provider as started on the plain
starting providerlog, so the next genericLLM request failedin this process is still attributed to Telegram even if no successful Bot API call ever happened. Gate Line 951 onreadyLoggedinstead so startup failures and post-ready inference failures stay distinct.Suggested fix
- if (providerStarted && /Embedded agent failed before reply|LLM request failed|FailoverError/i.test(text)) { + if (readyLogged && /Embedded agent failed before reply|LLM request failed|FailoverError/i.test(text)) { inferenceLogged = true; var line = text.split(/\r?\n/).find(function (entry) { return /Embedded agent failed before reply|LLM request failed|FailoverError/i.test(entry); }) || text; - emit('[telegram] [default] agent turn failed after provider startup; inference error: ' + sanitize(line).slice(0, 600)); + emit('[telegram] [default] agent turn failed after provider ready; inference error: ' + sanitize(line).slice(0, 600)); }🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@scripts/nemoclaw-start.sh` around lines 948 - 956, The current inference breadcrumb uses providerStarted to decide when to emit failures, which falsely attributes pre-ready LLM failures to Telegram; change the guard to require readyLogged instead of providerStarted (i.e., only set inferenceLogged and call emit(...) when readyLogged && /Embedded agent failed before reply|LLM request failed|FailoverError/i.test(text)); keep the same logic for extracting the matching line via text.split(...).find(...) and the sanitize(...).slice(0,600) call, and ensure any tests or downstream logic that relied on inferenceLogged still behave the same.
871-874:⚠️ Potential issue | 🟠 Major | ⚡ Quick winConsume the full Bearer token in
sanitize().Line 871 still stops at
:, and Line 873 only redacts up to the first whitespace afterauthorization:. A line likeAuthorization: Bearer 123456:ABCwill still leak the suffix. Please make the Bearer matcher consume the whole token, or foldBearer <token>into the authorization scrubber too.Suggested fix
- text = text.replace(/Bearer\s+[A-Za-z0-9._~+\/=-]+/g, 'Bearer <redacted>'); + text = text.replace(/Bearer\s+[^"'\s,)]+/gi, 'Bearer <redacted>'); text = text.replace( - /\b(api[_-]?key|token|authorization)\b(["']?\s*[:=]\s*["']?)[^"'\s,)]+/gi, + /\b(api[_-]?key|token|authorization)\b(["']?\s*[:=]\s*["']?)(?:Bearer\s+)?[^"'\s,)]+/gi, '$1$2<redacted>' );🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@scripts/nemoclaw-start.sh` around lines 871 - 874, The Bearer-redaction regex only matches a limited char set and stops at ":" (e.g. "Authorization: Bearer 123:ABC"), so update the sanitizer to consume the entire token by replacing the first regex with one that uses a non-whitespace match (for example use text.replace(/Bearer\s+\S+/gi, 'Bearer <redacted>')) or alternatively fold Bearer handling into the authorization/api key scrubber by expanding the second regex to also match "Bearer <token>" patterns; target the two occurrences shown (the line that calls text.replace(/Bearer\s+[A-Za-z0-9._~+\/=-]+/g, ...) and the authorization/api_key/token scrubber) and ensure the replacement uses '<redacted>' for the whole token.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Duplicate comments:
In `@scripts/nemoclaw-start.sh`:
- Around line 948-956: The current inference breadcrumb uses providerStarted to
decide when to emit failures, which falsely attributes pre-ready LLM failures to
Telegram; change the guard to require readyLogged instead of providerStarted
(i.e., only set inferenceLogged and call emit(...) when readyLogged && /Embedded
agent failed before reply|LLM request failed|FailoverError/i.test(text)); keep
the same logic for extracting the matching line via text.split(...).find(...)
and the sanitize(...).slice(0,600) call, and ensure any tests or downstream
logic that relied on inferenceLogged still behave the same.
- Around line 871-874: The Bearer-redaction regex only matches a limited char
set and stops at ":" (e.g. "Authorization: Bearer 123:ABC"), so update the
sanitizer to consume the entire token by replacing the first regex with one that
uses a non-whitespace match (for example use text.replace(/Bearer\s+\S+/gi,
'Bearer <redacted>')) or alternatively fold Bearer handling into the
authorization/api key scrubber by expanding the second regex to also match
"Bearer <token>" patterns; target the two occurrences shown (the line that calls
text.replace(/Bearer\s+[A-Za-z0-9._~+\/=-]+/g, ...) and the
authorization/api_key/token scrubber) and ensure the replacement uses
'<redacted>' for the whole token.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Enterprise
Run ID: c04366ed-2338-4a6b-a381-3ca55c5c022a
📒 Files selected for processing (4)
scripts/nemoclaw-start.shsrc/lib/onboard.tstest/nemoclaw-start.test.tstest/onboard.test.ts
✅ Files skipped from review due to trivial changes (1)
- test/nemoclaw-start.test.ts
Contributor
Selective E2E Results — ❌ Some jobs failedRun: 25226856804
|
Contributor
Selective E2E Results — ✅ All requested jobs passedRun: 25228189442
|
Contributor
Selective E2E Results — ✅ All requested jobs passedRun: 25226856804
|
…-2766' into fix/telegram-compatible-endpoint-2766
Contributor
Selective E2E Results — ✅ All requested jobs passedRun: 25229459569
|
Contributor
Selective E2E Results — ❌ Some jobs failedRun: 25232604443
|
Contributor
Author
the token-rotation-e2e is a known timeout, we can consider it passed for this purpose |
cv
approved these changes
May 1, 2026
12 tasks
cv
pushed a commit
that referenced
this pull request
May 2, 2026
## Summary Daily release-prep documentation refresh for merged PRs from the past 24 hours. This updates user-facing docs for Telegram mention-only mode, in-sandbox messaging shutdown, Hermes onboarding/runtime behavior, and compatible-endpoint smoke validation, then bumps the docs metadata to 0.0.33 after tag v0.0.32. ## Related Issue None. ## Changes - #2417 / c7e49ad: Document `TELEGRAM_REQUIRE_MENTION` for Telegram group-chat replies in `docs/manage-sandboxes/messaging-channels.md` and `docs/reference/commands.md`. - #1977 / 69403e0: Update `nemoclaw tunnel stop` and deprecated `nemoclaw stop` docs to explain that NemoClaw also attempts to stop the in-sandbox OpenClaw gateway and messaging polling. - #2781 / b83ffe2, #2859 / 4df8be6, and #2846 / 0968dfd: Refresh the Hermes quickstart for the default `my-hermes` sandbox name, cross-agent same-name guard, agent type visibility in `nemoclaw list`, Brave prompt omission, and supported prebaked Hermes integrations. - #2849 / fd240ff: Document the Telegram plus OpenAI-compatible endpoint `inference.local` smoke check in inference options and troubleshooting. - Bump `docs/versions1.json` and `docs/project.json` from 0.0.32 to 0.0.33 for daily release preparation. ## Type of Change - [ ] Code change (feature, bug fix, or refactor) - [ ] Code change with doc updates - [ ] Doc only (prose changes, no code sample modifications) - [x] Doc only (includes code sample changes) ## Verification - [ ] `npx prek run --all-files` passes - [ ] `npm test` passes - [ ] Tests added or updated for new or changed behavior - [x] No secrets, API keys, or credentials committed - [x] Docs updated for user-facing behavior changes - [x] `make docs` builds without warnings (doc changes only) - [x] Doc pages follow the [style guide](https://github.com/NVIDIA/NemoClaw/blob/main/docs/CONTRIBUTING.md) (doc changes only) - [ ] New doc pages include SPDX header and frontmatter (new pages only) Additional checks run: - `python3 scripts/docs-to-skills.py docs/ .agents/skills/ --prefix nemoclaw-user --dry-run` - `git diff --check` - `make docs` passed with the existing local version-switcher read message. - Full `npx prek run --all-files` and `npm test` were skipped for this doc-only automation run. Commit and pre-push hooks otherwise passed docs, lint, secret, and conversion checks until the local `Test (skills YAML)` hook failed because `vitest/config` is not installed in this fresh worktree. --- Signed-off-by: Miyoung Choi <miyoungc@nvidia.com> <!-- This is an auto-generated comment: release notes by coderabbit.ai --> ## Summary by CodeRabbit * **Documentation** * Updated Hermes quickstart: default sandbox name is "hermes"; guidance to use distinct sandbox names, note same-name reuse is prevented, Hermes wizard does not request Brave Web Search, and sandbox listings now show agent type. * Clarified provider onboarding: bounded in-sandbox smoke check runs when Telegram messaging is enabled. * Expanded Telegram docs: added TELEGRAM_REQUIRE_MENTION (DMs still governed by TELEGRAM_ALLOWED_IDS), onboarding examples, stop-messaging/tunnel behavior, and troubleshooting. * Promoted docs to version 0.0.33. <!-- end of auto-generated comment: release notes by coderabbit.ai --> --------- Signed-off-by: Miyoung Choi <miyoungc@nvidia.com> Co-authored-by: Cursor <cursoragent@cursor.com>
12 tasks
cv
added a commit
that referenced
this pull request
May 15, 2026
## Summary `NEMOCLAW_SANDBOX_READY_TIMEOUT` has been a recognised env var since #2849, but no documentation accompanied it — `docs/reference/commands.md`, `docs/reference/troubleshooting.md`, and the inference / deployment guides only mention the companion `NEMOCLAW_LOCAL_INFERENCE_TIMEOUT` (added in #1620 and documented at that time). Operators hitting `Sandbox '<name>' was created but did not become ready within 180s` have no doc-grep path to the workaround, and the two timeouts are easy to conflate. This closes the documentation gap left by #2849. Originally tried under #3435; closed because that PR mis-framed the docs as resolving #3344 / #3416 (the root cause of both was the GPU policy bug fixed in #3436, not a timeout misconfiguration). The docs themselves still have value as a follow-up to the env-var introductions, so reopening as a new PR with the correct framing. ## Related Issue <!-- Not closing any issue; this addresses the doc-gap surfaced while investigating #3344 and #3416 (both already fixed in code by #3436). --> ## Changes - `docs/reference/commands.md`: add `NEMOCLAW_SANDBOX_READY_TIMEOUT` and `NEMOCLAW_LOCAL_INFERENCE_TIMEOUT` to the Onboard Timeouts table. - `docs/reference/troubleshooting.md`: new troubleshooting entry "Sandbox onboard times out with 'did not become ready within Ns'" that distinguishes the readiness wait from the inference-probe budget, with a worked example. - `docs/inference/use-local-inference.md`: cross-link the two timeouts from the existing `NEMOCLAW_LOCAL_INFERENCE_TIMEOUT` section so readers of either knob land on the other. - `docs/deployment/deploy-to-remote-gpu.md`: new "First-Run Readiness Budget" section calling out DGX Station / cloud-VM / large-quantised-model conditions that exceed the default and showing how to raise it. No code changes — the readiness behaviour is unchanged. ## Type of Change - [ ] Code change (feature, bug fix, or refactor) - [ ] Code change with doc updates - [x] Doc only (prose changes, no code sample modifications) - [ ] Doc only (includes code sample changes) ## Verification - [ ] `npx prek run --all-files` passes - [ ] `npm test` passes - [ ] Tests added or updated for new or changed behavior - [ ] No secrets, API keys, or credentials committed - [ ] Docs updated for user-facing behavior changes - [x] `make docs` builds without warnings (doc changes only) - [x] Doc pages follow the [style guide](https://github.com/NVIDIA/NemoClaw/blob/main/docs/CONTRIBUTING.md) (doc changes only) - [ ] New doc pages include SPDX header and frontmatter (new pages only) --- Signed-off-by: Tinson Lai <tinsonl@nvidia.com> <!-- This is an auto-generated comment: release notes by coderabbit.ai --> ## Summary by CodeRabbit * **Documentation** * Added a “First-Run Readiness Budget” note for remote GPU hosts explaining longer initial sandbox build/upload times and advice to increase NEMOCLAW_SANDBOX_READY_TIMEOUT. * Clarified that NEMOCLAW_LOCAL_INFERENCE_TIMEOUT applies to inference-server validation while sandbox readiness uses NEMOCLAW_SANDBOX_READY_TIMEOUT (default 180s). * Expanded examples for exporting both timeouts and onboarding timeout messaging. * Added troubleshooting guidance and inspection steps when sandbox readiness timeouts delete partial sandboxes. <!-- review_stack_entry_start --> [](https://app.coderabbit.ai/change-stack/NVIDIA/NemoClaw/pull/3440) <!-- review_stack_entry_end --> <!-- end of auto-generated comment: release notes by coderabbit.ai --> --------- Signed-off-by: Tinson Lai <tinsonl@nvidia.com> Co-authored-by: Carlos Villela <cvillela@nvidia.com>
](https://app.coderabbit.ai/change-stack/NVIDIA/NemoClaw/pull/3440){kind=link}
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Adds a targeted guard for the Telegram plus OpenAI-compatible endpoint regression so onboarding validates the managed inference.local route before a messaging-triggered agent turn fails later with a generic network error.
Related Issue
Fixes #2766
Changes
https://inference.local/v1/chat/completions.messaging-compatible-endpoint-e2ebranch-validation job with a local OpenAI-compatible mock endpoint.Type of Change
Verification
npx prek run --all-filespassesnpm testpassesmake docsbuilds without warnings (doc changes only)Signed-off-by: Aaron Erickson aerickson@nvidia.com
Summary by CodeRabbit
New Features
Tests
Chores
Documentation