github modelcontextprotocol/registry v1.8.5

5 hours ago

This release adds production performance diagnostics: request arrivals and backlog, database pool/query metrics, and bounded slow-listing logs correlated with ingress request IDs. Slow-request logs separate connection acquisition, query/fetch waits, row decoding, and response processing, including snapshots while requests remain stuck.

It also includes publisher fixes for P-384 signatures and repository IDs, GitLab subgroup URL support, and stricter validation of incomplete repository metadata and invalid arguments. Previously stored invalid metadata may need correction when republishing or editing.

The ingress cache/timing log fields require the accompanying infrastructure rollout. Application release artifacts alone do not deploy the hosted registry. No new database migration is included.

What's Changed

  • chore(deploy): promote production to 1.8.4 by @rdimitrov in #1732
  • fix(publisher): pad ECDSA P-384 signature to fixed width by @JosephDoUrden in #1457
  • fix(validators): reject incomplete repository metadata by @samrusani in #1555
  • fix(validators): allow GitLab repository URLs with nested subgroups by @anneheartrecord in #1361
  • fix(validators): reject invalid argument types and empty positional arguments by @JosephDoUrden in #1583
  • fix(publisher): record repository.id at init so renames stay resolvable by @UgaTheDev in #1570
  • docs(security): pin and verify mcp-publisher installer by @sys-ops-node in #1523
  • docs(auth): document HTTP verification requirements and pin them with tests by @ump45nose in #1640
  • docs(publishing): troubleshooting for DNS/HTTP auth (#845) by @rhinocap in #1268
  • docs(api): clarify per-version listings and latest filtering by @rdimitrov in #1735
  • docs: add ProtocolProbe to community projects by @Sultan57git in #1736
  • fix(deploy): reduce ingress restarts under load by @rdimitrov in #1737
  • deploy: promote ingress stability fixes to production by @rdimitrov in #1738
  • deploy: upgrade GKE boot disks to 50 GB balanced storage by @rdimitrov in #1739
  • deploy: promote balanced node disks to production by @rdimitrov in #1740
  • feat(observability): expose request backlog and database waits by @rdimitrov in #1742
  • feat(observability): diagnose individual slow listing requests by @rdimitrov in #1743

New Contributors

Full Changelog: v1.8.4...v1.8.5

Don't miss a new registry release

NewReleases is sending notifications on new releases.