DevOps

Monitor and troubleshoot GKE TPU Dynamic Slices and their lifecycle states

Monitor GKE TPU duty cycle, memory, node readiness and interruptions with PromQL

Diagnose multi-slice TPU training hangs using Megascale hang logs and monitored events

Resolve GKE TPU slice creation failures caused by overlapping partitions

Diagnose GKE TPU throughput drops and step-time regressions with monitoring data

Resolve GKE TPU node pool scale-up failures caused by reservation stockouts

Fix TPU slice and JobSet failures blocked by instance groups stuck updating

Prevent vbar control agent crashes and OOM errors on GKE TPU v6e nodes

Diagnose GKE TPU VMs stuck in Compute Engine repair after host errors

Write Terraform alerting policies for GKE latency, traffic and errors using PromQL

Configure Backup for GKE with backup plans, restore plans and CMEK encryption

Configure GKE ComputeClasses for Spot VMs, GPUs and on-demand fallback

Analyze GKE costs by project, namespace and workload from BigQuery billing

Rightsize GKE workloads and pick Spot VMs and committed use discounts

Find the golden base image for a custom GKE node image from cluster specs

Check GKE clusters against production readiness and golden path defaults

Deploy LLM inference servers on GKE with GPU or TPU resources

Generate secure Kubernetes YAML manifests for GKE Autopilot and Standard

Plan GKE multi-tenancy with namespaces, RBAC, quotas and network isolation

Plan and configure GKE networking, private clusters and Dataplane V2

Diagnose GKE nodes stuck in NotReady status and propose safe fixes

Set up Cloud Logging, Cloud Monitoring and managed Prometheus for GKE

Harden GKE cluster security with RBAC, Binary Authorization and encryption

Run a GKE production readiness review across security, reliability and cost

Improve GKE workload reliability with PDBs, health probes and spread constraints

Configure GKE Gateway API, Ingress, Cloud Armor and load balancing

Configure GKE persistent volumes, PVCs and Filestore storage

Diagnose GKE volume attach, mount and disk performance failures

Plan and run GKE cluster upgrades with checklists and maintenance runbooks

Fix GKE Workload Identity auth failures for Pods and service accounts

Configure HPA and VPA autoscaling for GKE workloads

Diagnose GKE HPA failures, unknown metrics and broken metrics pipelines

Audit and harden GKE workload security with network policies and sandboxing

Diagnose GKE pod failures like CrashLoopBackOff, OOMKilled and Pending

Create PromQL SLO alerting policies in Cloud Monitoring with Terraform output

Migrate Cloud Run or Gemini API AI workloads to self-hosted GKE inference

Triage on-call alerts with Sentry stack traces, Datadog logs and recent deploy checks

Draft postmortems for resolved incidents from channel timelines, deploys and log spikes

Configure a managed or self-supervised iii engine with ports, RBAC and storage

Trace, log and alert on engine telemetry with OpenTelemetry metrics and sampling

Configure Windmill schedules that run scripts and flows on a timer

Read and change Windmill workspace settings in the settings.yaml that wmill sync uses

Write Ansible playbooks as Windmill scripts for infrastructure automation

Write Ansible playbooks, roles and inventories for infrastructure automation

Build CI/CD pipelines in GitHub Actions, GitLab CI and Jenkins

Build Docker images, Compose stacks and containers and debug failures

Package Kubernetes apps with Helm charts, templates and dependencies

Operate Kubernetes with kubectl, deployments and pod debugging

Configure Nginx for reverse proxies, load balancing, TLS and tuning