Repository navigation
fix(vmcp): allow server/discover in peer health probes (PAIINFRA-399) - #28
Merged
carlos-lopez-vecino merged 1 commit intoOct 26, 2026
Merged
Conversation
Add server/discover to peerProbeMethods so Modern-protocol health checks from a hub are not refused with 403. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
jgpelaez
approved these changes
Oct 26, 2026
davidrubio-dp
approved these changes
Oct 26, 2026
carlos-lopez-vecino
deleted the
PAIINFRA-399/peer-probe-server-discover
branch
October 26, 2026 09:50
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Problem
After staging moved to v0.42.1 (DocPlanner/platform-apps#2047), the staging hubs lost the
i0andglobalbackends. The tool count dropped from 698 to 154.Root cause
Fork patch 43623c1 (
peer_probe.go) lets health-probe requests (headerX-ToolHive-Health-Check) use onlyinitialize,pingand the*/listmethods.v0.42.1 added the Modern (2026-07-28) protocol. Once a hub caches a backend as Modern, its health checker opens with
server/discover. The backend vMCP refuses that with 403.The hub maps the 403 to
errModernAuth, then toErrAuthenticationFailed. Backends with no outgoing auth (in-cluster i0 and global) are markedunauthenticatedbyauthErrorStatus. The circuit breaker opens and their tools drop out of aggregation.Evidence:
server/discovercalls (0 ms, empty clientCapabilities, so the health checker).Side effect
istanbul and security also get the 403. They have outgoing auth, so
authErrorStatuscounts the 403 as Healthy. Their health checks have been hollow since v0.42.1. This fix covers them too.Upstream
Upstream is not affected. Neither v0.42.1 nor upstream main reads the probe header server-side. This is a fork-only bug. The v0.42.1 merge (#27) missed updating the allowlist.
Fix
Add
server/discovertopeerProbeMethods, with a comment.There is no new exposure.
server/discoveris already always allowed without the probe header (pkg/authz/middleware.go).Testing
server/discoveris passed and marked.subscriptions/listenstays refused.go test ./pkg/vmcp/server/passes (65 s).go vetis clean.Rollout
vmcpImageoftoolhive-operator-stgin platform-appsapp-of-apps/chart/values-platform.yaml. Bump the other images too, for consistency.Production (PAIINFRA-242) must include this fix before it moves to v0.42.1.
🤖 Generated with Claude Code