All pull requests

Pull requests

• feat(llm-routing): deploy independent models through a shared stack

  #2340·priyaselvaganesan opened Oct 7, 2026Contributor·4 tasks done·

  ·
  1

• feat(nvcf-cli): forward routing expressions unchanged

  #2339·along-2017 opened Oct 7, 2026Contributor·4 tasks done·

  ·
  2

  3

• fix(event-ledger): authorize JWTs with route scopes

  #2329·borao opened Oct 6, 2026Contributor·4 tasks done·

  ·
  1

  6

• feat(vanity-gateway): label the origin of non-2xx responses

  #2328·FamousDirector opened Oct 6, 2026Contributor·4 tasks done·

  ·
  1

  4

• test(stargate): drive deployed fleets with the routing simulator workload

  #2321·barrygreengus opened Oct 6, 2026Contributor·5/5·4 tasks done·

  ·
  24

• feat(nvsnap): checkpoint and restore tensor-parallel workloads through gpushare

  #2318·balajinvda opened Oct 6, 2026Contributor·

  ·
  3

• fix(openbao): scope the init-cluster Job Secret RBAC by name

  #2317·sunilthorat09 opened Oct 6, 2026·

  ·
  1

  3

• fix(pylon): judge canary runaway by exact usage, not character estimates

  #2315·FamousDirector opened Oct 6, 2026Contributor·4 tasks done·

  ·
  1

  4

• docs(release-notes): make the 0.6.2 to 1.0.1 upgrade guide runnable e…

  #2312·Manasi-NV opened Oct 6, 2026Contributor·

  ·
  4

• fix(nvcf-cli): fail self-hosted check --control-plane and --compute-plane instead of reporting a vacuous pass

  #2311·rohithb-hub opened Oct 6, 2026Contributor·4 tasks done·

  ·
  3

• feat(openbao): support KMS auto-unseal in the self-managed chart

  #2310·sunilthorat09 opened Oct 6, 2026·

  ·
  1

  29

• feat(nvsnap): import gpushare, checkpoint support for GPU memory shared between processes

  #2300·balajinvda opened Oct 6, 2026Contributor·4 tasks done·

  ·
  51

• fix(grpc-proxy): bound concurrent HTTP/2 streams

  #2297·efegokdemir opened Oct 5, 2026·4 tasks done·

  ·
  1

  1

• fix(pylon): exclude cached prompt tokens from fallback TPS

  #2296·efegokdemir opened Oct 5, 2026·4 tasks done·

  ·
  1

  5

• fix(stargate): expire reservations after backend RTT

  #2295·efegokdemir opened Oct 5, 2026·4 tasks done·

  ·
  1

• fix(cloud-functions): reject llmConfig.uris and tokenizer in modelUpdates

  #2292·Max-NV opened Oct 5, 2026Contributor·

  ·
  1

  11

  

• feat(stargate): make pulsar-wait-and-widen wait for affinity and widen gradually

  #2288·barrygreengus opened Oct 5, 2026Contributor·2/5·4 tasks done·

  ·
  1

  7

• feat(mock-dynamo): model batched GPU workers with a shared engine crate

  #2286·barrygreengus opened Oct 5, 2026Contributor·3/5·4 tasks done·

  ·
  1

  9

• fix(nvca): set up RBAC before writing the agent config ConfigMap

  #2283·apartha-nv opened Oct 5, 2026Contributor·

  ·
  1

  1

• docs(self-managed): point the 1.0.1 sample-file links at their current path

  #2263·balajinvda opened Oct 5, 2026Contributor·

  ·
  1

  2

• feat(event-ledger): add cache, timing wheel, and metrics

  #2255·adi0106 opened Oct 3, 2026Contributor·2 of 4 tasks·

  ·
  1

  12

• fix(nvca-operator): honor --system-namespace in BackendK8sCache

  #2250·cestercian opened Oct 3, 2026Contributor·4 tasks done·

  ·
  1

  6

• fix(nvca-operator): derive sibling image repositories from image.repository

  #2238·balajinvda opened Oct 2, 2026Contributor·

  ·
  1

  1

• fix(nvca): size the model cache claim with headroom and report a full cache volume

  #2234·balajinvda opened Oct 2, 2026Contributor·

  ·
  1

• fix(openbao): remove Helm from the migration runtime and hook checks

  #2232·sbaum1994 opened Oct 2, 2026Collaborator·

  ·
  4