Cloud, CI/CD, monitoring, and deployment (6.8K skills)
Deploy and tune DOCA's supplied ZTR RTTCC algorithm on BlueField-3 to adjust RoCE-v2 traffic under congestion. Get guidance on sample integration, build-time variants, parameter changes, and diagnosing cases where the loaded algorithm appears inactive.
Guides initial BlueField-4 setup through its BMC, the hardware management controller. Covers OS installation methods, firmware updates, post-install checks, and troubleshooting, with action-specific confirmation and maintenance safeguards for changes that can disrupt or damage the device.
Bring up BlueField-3 hardware through RShim, its host management interface, and BFB system images. Follow installation and recovery workflows, diagnose management connectivity and boot state, and verify the installed system and version compatibility.
Analyze live or captured DOCA Flow pipelines to guide changes to rule placement, table sizing, and hardware offloading. The skill helps interpret snapshots, compare measurements before and after changes, and investigate missing or inconsistent tuning results.
Use pcc_counters.sh to arm and read fixed congestion-control diagnostic counters on ConnectX or BlueField devices. Capture CNP, RTT, and WRED-drop readings, and investigate invalid device paths, empty output, or counters that remain at zero.
Build and run the GPUNetIO benchmark to measure sustained RDMA WRITE bandwidth, a remote-memory transfer rate driven by a CUDA kernel. Interpret results, investigate run-to-run variation, and identify GPU, network adapter, or link bottlenecks.
Measure throughput, latency, or maximum bandwidth for DOCA libraries using doca_bench on a host or BlueField Arm. Capture commands and environment details for repeatable comparisons, check run stability, and investigate configuration or measurement problems.
Deploy and troubleshoot a CollectX telemetry collector on a host or BlueField. Connect sources of device metrics, configure outputs such as Prometheus or Fluent Bit, and verify that collected measurements reach their downstream destination.
Connect an existing socket application to a BlueField DPU without rewriting the application. This skill guides relay placement, endpoint configuration, a single-client round-trip check, and diagnosis of refused connections or traffic that stops flowing.
Operate NVIDIA DOCA Management Service to read, change, and monitor BlueField or ConnectX settings locally or remotely. Get guidance on deployment, authentication, management requests, logs, and diagnosing failures across the service and its backend.
Run DOCA service containers on BlueField using YAML deployment files and standalone kubelet, the agent that starts containers. Check service readiness and troubleshoot startup failures, image downloads, missing mounts, and repeated restarts.
Set up and troubleshoot a gRPC server, a service for remote procedure calls, to control DOCA Flow from another host or languages such as Python. Covers connection checks, interface definitions, and external transport protection.
Measure how quickly a host or DPU CPU installs and deletes DOCA Flow rules. Choose a JSON workload and backend, run repeated benchmarks, and record hardware, software, and execution settings so results can be assessed.
Prepare changes to live BlueField or ConnectX hardware with baseline records, a separate management connection, a maintenance window, and rehearsal on matching test hardware. Define verification and rollback steps before applying firmware or configuration changes.
Run and supervise DOCA applications directly on an x86 host or BlueField Arm hardware. Covers process launch, hardware resource assignment, isolation between workloads, and troubleshooting application failures or BlueField recovery problems.
Helps inspect comch communication channels between a host and a BlueField DPU using the DOCA admin tool. Explains server and connection tables and checks discrepancies with application state through a read-only inventory scan.
Configure and troubleshoot DOCA Firefly time synchronization on BlueField. The skill helps decide whether high-precision PTP synchronization is needed, choose service settings, and check that the host clock follows BlueField's hardware clock.
Inspect which devices, libraries, and capabilities DOCA recognizes on a host using a read-only command-line tool. Narrow results to a PCIe address and capture a capability snapshot for installation checks or troubleshooting.
Install cuOpt for Python, C, or use as a REST server, a service called over HTTP. Choose pip, conda, or Docker to match the environment, verify installation, and troubleshoot package or CUDA mismatches.
Set up AutoMagicCalib from NGC release containers using Docker Compose. The skill handles registry authentication, port and path configuration, optional VGGT model download, service startup, and readiness checks for the backend and web interface.
Prepare the KERMT environment by checking Docker and GPU access, building the local container image when needed, and testing whether PyTorch detects GPUs inside it. Report the image details and detected GPU count before other KERMT workflows begin.
Review deployment status across projects using ECS service data, Vercel deployment history, and GitHub Actions runs. Get a consolidated dashboard of running services, recent releases, pipeline failures, and changes waiting to be deployed.
Check the claude-ops background process for outdated paths, stopped services, configuration problems, and incompatible Bash versions. Review proposed repairs individually, apply approved fixes, and run checks again to confirm the resulting state.
Review integration credentials after initial setup, check which services are configured or missing, and update individual keys. Connection checks and post-update smoke tests help reveal credentials that are saved but fail to work.