Full Changelog: v1.13.0...v1.13.1

What's Changed

New Features

• feat(cluster): enable GKE managed ML Diagnostics for GKE version >= 1.35.0-gke.3065000 by @rapatchi in #1186
• Upgrade Kueue and JobSet by @jamOne- in #1189

Improvements

• build: pin google-cloud-bigquery>=3.10.0 in pyproject.toml by @rapatchi in #1185
• Add a adapt_from_ct cluster create flag by @SikaGrr in #1188
• Remove CLUSTER_NETWORK_ARGUMENTS_DWS variable by @jamOne- in #1190

Bug fixes

• feat: move colocated-python-sidecar to worker job in pathways template by @ChromeHearts in #1182
• FIx scheduling of nap + pathways workloads by @scaliby in #1184
• Grant required IAM roles to Compute Engine default SA when --managed-mldiagnostics is passed during xpk cluster create by @rapatchi in #1187
• fix: Fix gcluster and make Nightly Tests passing again by @jamOne- in #1191

New Contributors

• @ChromeHearts made their first contribution in #1182
• @rapatchi made their first contribution in #1185

Full Changelog: v1.12.0...v1.13.0

What's Changed

Improvements

• Update kubernetes dependency to version 35.0.0 by @jamOne- in #1178

Full Changelog: v1.11.0...v1.12.0

What's Changed

New Features

• feat: support --device-type for xpk cluster create-ray by @jamOne- in #1175

Full Changelog: v1.10.0...v1.11.0

What's Changed

Improvements

• Clarify path installation is only for installing from source by @bvandermoon in #1170

Bug fixes

• Fix circular import issue by @FIoannides in #1166

New Contributors

• @bvandermoon made their first contribution in #1170

Full Changelog: v1.9.0...v1.10.0

What's Changed

New Features

• Add support for n2-standard-4 machine type. by @cpgaffney1 in #1128
• feat: Support Super-slicing Pathways workload by @jamOne- in #1119
• Add NATIVE_CLUSTER_TOOLKIT_ENABLED feature flag by @scaliby in #1133
• Add gcluster binary to Dependency Auto Download (DAD) mechanism by @scaliby in #1137
• feat: --no-use-parallel-containers in workload create by @jamOne- in #1136
• Add --dependency-auto-download to gcluster executions by @scaliby in #1139
• feat: add private control plane endpoint support with custom subnet by @kryvokhyzha in #1124
• Launch Native Cluster Toolkit Execution by @scaliby in #1158

Improvements

• chore: Migrate gsutil usage to gcloud storage by @gurusai-voleti in #1064
• Add consent to delete old workload on duplicate by @scaliby in #1126
• Improve workload scheduling error message by @scaliby in #1131
• Extract CommandRunner base class to its own file by @scaliby in #1138
• Fix mtc updates by @FIoannides in #1115
• Increase kueue_manager cpu and memory by @SikaGrr in #1140
• Fix variable name for super slicing check by @jamOne- in #1142
• Move deployment dir computation to CommandRunner by @scaliby in #1145
• Implement NativeCommandRunner by @scaliby in #1146
• Update check for topology requirement by @FIoannides in #1150
• Improve error message for host-maintenance-interval error by @scaliby in #1153
• Update cluster toolkit + fix zip downloads by @scaliby in #1156
• Remove docker as a prerequisite for using xpk in docs by @scaliby in #1159

Bug fixes

• fix: correct placement of jobset exclusive placement annotations for Pathways by @jamOne- in #1122
• Fix formatting of Pathways workload message by @jamOne- in #1129
• ci: remove caching from GitHub Actions to deflake presubmits by @jamOne- in #1127
• fix: Update placement policy condition to include reservation by @jamOne- in #1130
• fix: xpk workload list Accelerator VMs counting by @jamOne- in #1132
• gcluster DAD flag by @scaliby in #1144
• fix: stacktrace-explorer sidecar hangs after main container exits by @kryvokhyzha in #1134
• fix(workload): resolve Kubernetes v0.15+ condition sorting issue by @jamOne- in #1149
• Fix --num-slices fallback calculation by @scaliby in #1152
• Update zone for GPU test by @FIoannides in #1100
• Update jobset used by GPUs from version 0.7.2 to 0.8.1 by @FIoannides in #1151
• Fix master ipv4 cidr psc by @FIoannides in #1154
• Fix inspector slice logs by @scaliby in #1155
• fix: JobSet Status Resolution by @jamOne- in #1161

New Contributors

• @gurusai-voleti made their first contribution in #1064
• @cpgaffney1 made their first contribution in #1128

Full Changelog: v1.8.0...v1.9.0

What's Changed

New Features

• Pathwaysjob CRD migration by @FIoannides in #1099
• Enable Crane by @SikaGrr in #1120

Improvements

• Update docs to accomodate for DAD by @scaliby in #1102
• Refactoring workload list parsing to Python by @jamOne- in #1089
• Refactor pathways workload scheduling to use Jinja template by @jamOne- in #1116

Bug fixes

• fix: workload list priorityClassName for super-slicing workloads by @jamOne- in #1105
• Revert "Pathwaysjob CRD migration (#1099)" by @jamOne- in #1107
• fix: xpk workload list should sum podSet counts by @jamOne- in #1106
• Pathwaysjob migration by @FIoannides in #1108
• fix aggregate reservation accelerator type matching for DWS Calendar by @kryvokhyzha in #1114
• fix: relax Kueue version check for sub and super-slicing workloads by @jamOne- in #1111
• fix: skip reservation capacity assessment when no nodepools need to be created by @jamOne- in #1125

New Contributors

• @kryvokhyzha made their first contribution in #1114

Full Changelog: v1.7.0...v1.8.0

What's Changed

New Features

• Add DEPENDENCY_AUTO_DOWNLOAD flag by @scaliby in #1088
• Add script to compute dependencies checksums by @scaliby in #1091
• Add dependency auto download downloader logic by @scaliby in #1094
• Integrate deps downloading with the rest by @scaliby in #1095
• Launch DAD + Update dependencies by @scaliby in #1101

Improvements

• Align download and install directory by @wstcliyu in #1090
• Add .multi_agent to .gitignore by @jamOne- in #1092
• feat: add flag to disable DAD by @scaliby in #1097
• Ignore nodepool creation errors by @scaliby in #1096

Bug fixes

• fix: Limit super-slicing workload name by @jamOne- in #1087

Full Changelog: v1.6.0...v1.7.0

What's Changed

New Features

• Reservations capacity assesment by @jamOne- in #1057
• feat: add tpu7 support by @bhuvanpkaruturi in #1072
• feat: allow Numa aware workloads with super-slicing by @jamOne- in #1080
• Add customizable binaries support by @scaliby in #1085

Improvements

• Ignore python venv better by @jamOne- in #1083

Bug fixes

• Add TELEMETRY_TRASH_EXECUTION to goldens and verify-goldens by @jamOne- in #1071
• Fix b/485965692: Support custom topologies in superslicing by @jamOne- in #1082

New Contributors

• @bhuvanpkaruturi made their first contribution in #1072

Full Changelog: v1.5.0...v1.6.0

What's Changed

New Features

• feat: Increase number of concurrent NPs to be created to 100 by @jamOne- in #1068
• docs: Add Super-slicing docs by @jamOne- in #1069

Improvements

• Parallelize make goldens execution by @jamOne- in #1061
• make verify-goldens by @jamOne- in #1062
• Add crane workload integration test by @SikaGrr in #1063
• fix: Super-slicing cluster documentation by @jamOne- in #1070

Full Changelog: v1.4.0...v1.5.0