Nomad

Nomad Skill

Orchestrate workloads with the HashiCorp Nomad CLI.

Requirements

• nomad CLI installed
• NOMAD_ADDR environment variable (default: http://127.0.0.1:4646)
• NOMAD_TOKEN if ACLs are enabled

Jobs

List & Status

nomad job status                      # List all jobs
nomad job status <job-id>             # Job details
nomad job status -verbose <job-id>    # Detailed info

Run & Stop

nomad job run job.nomad.hcl           # Deploy job
nomad job run -check-index=X job.hcl  # Deploy only if index matches
nomad job plan job.nomad.hcl          # Dry-run, show changes
nomad job stop <job-id>               # Stop job (leaves history)
nomad job stop -purge <job-id>        # Stop and remove completely

Scale

nomad job scale <job-id> <group> 5    # Scale to 5 instances
nomad job scale -detach <job-id> <group> 3

Dispatch (Parameterized Jobs)

nomad job dispatch <job-id>
nomad job dispatch -meta key=value <job-id>
nomad job dispatch -meta key=value <job-id> @payload.json

History & Versions

nomad job history <job-id>
nomad job revert <job-id> <version>
nomad job deployments <job-id>

Inspect

nomad job inspect <job-id>            # Full job spec JSON
nomad job validate job.nomad.hcl      # Validate syntax

Allocations

nomad job allocs <job-id>             # List job allocations
nomad alloc status <alloc-id>         # Allocation details
nomad alloc status -short <alloc-id>  # Summary only

Logs

nomad alloc logs <alloc-id>           # stdout
nomad alloc logs -stderr <alloc-id>   # stderr
nomad alloc logs -f <alloc-id>        # Follow
nomad alloc logs -tail -n 100 <alloc-id>
nomad alloc logs -job <job-id>        # Latest alloc logs
nomad alloc logs <alloc-id> <task>    # Specific task

Exec

nomad alloc exec <alloc-id> /bin/sh
nomad alloc exec -task <task> <alloc-id> /bin/sh
nomad alloc exec -i -t <alloc-id> /bin/bash

Files

nomad alloc fs <alloc-id>                    # List /
nomad alloc fs <alloc-id> /alloc/logs/       # List directory
nomad alloc fs -stat <alloc-id> /path        # File metadata
nomad alloc fs -tail -f <alloc-id> /alloc/logs/app.stdout.0

Actions

nomad alloc stop <alloc-id>           # Stop allocation
nomad alloc restart <alloc-id>        # Restart all tasks
nomad alloc restart -task web <alloc-id>  # Restart specific task
nomad alloc signal -s SIGHUP <alloc-id>   # Send signal

Nodes

nomad node status                     # List nodes
nomad node status <node-id>           # Node details
nomad node status -allocs <node-id>   # Show allocations
nomad node status -verbose <node-id>  # Full details

Node Management

nomad node drain -enable <node-id>    # Drain for maintenance
nomad node drain -disable <node-id>   # Stop draining
nomad node eligibility -enable <node-id>   # Allow scheduling
nomad node eligibility -disable <node-id>  # Prevent scheduling

Deployments

nomad deployment list
nomad deployment status <deployment-id>
nomad deployment promote <deployment-id>   # Promote canary
nomad deployment fail <deployment-id>      # Mark failed
nomad deployment pause <deployment-id>
nomad deployment resume <deployment-id>

Evaluations

nomad eval list
nomad eval status <eval-id>

Services

nomad service list
nomad service info <service-name>
nomad service delete <service-name>   # Force delete

Namespaces

nomad namespace list
nomad namespace status <name>
nomad namespace apply -description "Prod" production
nomad namespace delete <name>

Variables

nomad var list
nomad var get <path>
nomad var put <path> key=value key2=value2
nomad var put <path> @data.json
nomad var purge <path>

ACLs

nomad acl policy list
nomad acl policy info <name>
nomad acl token list
nomad acl token self                  # Current token info

Cluster

nomad server members                  # Raft cluster
nomad operator raft list-peers        # Raft peers
nomad agent-info                      # Local agent info
nomad operator snapshot save backup.snap
nomad operator snapshot restore backup.snap

JSON Output

Add -json to most commands:

nomad job status -json
nomad node status -json
nomad alloc status -json <alloc-id>

Filtering

Use -filter with expressions:

nomad job status -filter='Status == "running"'
nomad node status -filter='Status == "ready"'
nomad job allocs -filter='ClientStatus == "failed"' <job-id>

Common Patterns

Logs from latest allocation

nomad alloc logs $(nomad job allocs -json <job> | jq -r '.[0].ID')

Find failed allocations

nomad job allocs <job-id> | grep -i failed
nomad eval list -filter='Status == "blocked"'

Quick restart

nomad job stop <job-id> && nomad job run job.nomad.hcl

Force new deployment

nomad job run -force job.nomad.hcl

Exec into running container

ALLOC=$(nomad job allocs -json <job> | jq -r '.[0].ID')
nomad alloc exec -i -t $ALLOC /bin/sh

Watch deployment

nomad deployment status -monitor <deployment-id>

Scale to zero and back

nomad job scale <job-id> <group> 0
nomad job scale <job-id> <group> 3

Environment Variables

• NOMAD_ADDR — API address (default: http://127.0.0.1:4646)
• NOMAD_TOKEN — ACL token
• NOMAD_NAMESPACE — Default namespace
• NOMAD_REGION — Default region
• NOMAD_CACERT / NOMAD_CLIENT_CERT / NOMAD_CLIENT_KEY — TLS config
• NOMAD_TLS_SERVER_NAME — TLS server name for verification

Tips

• Use nomad-tui for interactive cluster management
• Use -monitor flag with deployments to watch progress
• Prefix alloc IDs work (e.g., nomad alloc logs abc123 if unique)
• -detach returns immediately without waiting
• -verbose adds scheduling info and constraint details