managing-nomad

name	managing-nomad
description	Use when working with Nomad — hashiCorp Nomad workload orchestrator management. Covers job submission, allocation status, deployment health, client/server status, namespace management, and resource utilization. Use when managing Nomad jobs, debugging allocation failures, monitoring deployments, or inspecting cluster health.
connection_type	nomad
preload	false

Nomad Management Skill

Manage HashiCorp Nomad jobs, allocations, deployments, and cluster health.

Core Helper Functions

#!/bin/bash

# Nomad CLI wrapper
nomad_cmd() {
    nomad "$@" 2>/dev/null
}

# Nomad HTTP API helper
nomad_api() {
    local method="${1:-GET}"
    local endpoint="$2"
    local data="${3:-}"
    local url="${NOMAD_ADDR:-http://127.0.0.1:4646}"

    local auth_header=""
    if [ -n "${NOMAD_TOKEN:-}" ]; then
        auth_header="-H \"X-Nomad-Token: $NOMAD_TOKEN\""
    fi

    if [ -n "$data" ]; then
        eval curl -s -X "$method" $auth_header -H "Content-Type: application/json" "${url}/v1/${endpoint}" -d "'$data'"
    else
        eval curl -s -X "$method" $auth_header "${url}/v1/${endpoint}"
    fi
}

MANDATORY: Discovery-First Pattern

Always discover cluster status, nodes, and jobs before performing specific operations.

Phase 1: Discovery

#!/bin/bash

echo "=== Nomad Agent Info ==="
nomad_api GET "agent/self" | jq '{
    name: .member.Name,
    region: .config.Region,
    datacenter: .config.Datacenter,
    version: .config.Version,
    server: .config.Server
}'

echo ""
echo "=== Server Members ==="
nomad_api GET "agent/members" | jq -r '.Members[] | "\(.Name)\t\(.Addr):\(.Port)\t\(.Status)\t\(.Tags.region)\t\(.Tags.dc)"' | column -t

echo ""
echo "=== Client Nodes ==="
nomad_api GET "nodes" | jq -r '.[] | "\(.ID[0:8])\t\(.Name)\t\(.Status)\t\(.Drain)\t\(.NodeClass // "default")\t\(.Datacenter)"' | column -t | head -20

echo ""
echo "=== Job Summary ==="
nomad_api GET "jobs" | jq -r '.[] | "\(.ID)\t\(.Type)\t\(.Status)\t\(.Namespace // "default")\t\(.Priority)"' | column -t | head -30

Output Rules

• TOKEN EFFICIENCY: Target <=50 lines per output
• Use Nomad HTTP API with jq for structured output
• Always use short allocation IDs (first 8 chars) for readability

Common Operations

Job Management Dashboard

#!/bin/bash
echo "=== Job Status Summary ==="
nomad_api GET "jobs" | jq '{
    total: length,
    running: [.[] | select(.Status == "running")] | length,
    dead: [.[] | select(.Status == "dead")] | length,
    pending: [.[] | select(.Status == "pending")] | length,
    by_type: (group_by(.Type) | map({type: .[0].Type, count: length}))
}'

echo ""
echo "=== Failed Jobs ==="
nomad_api GET "jobs" | jq -r '
    .[] | select(.Status == "dead") | "\(.ID)\t\(.Type)\t\(.StatusDescription[0:60])"
' | column -t | head -10

echo ""
echo "=== Recent Deployments ==="
nomad_api GET "deployments" | jq -r '
    sort_by(.CreateIndex) | reverse | .[:10][] |
    "\(.ID[0:8])\t\(.JobID)\t\(.Status)\t\(.StatusDescription[0:40])"
' | column -t

Allocation Debugging

#!/bin/bash
JOB="${1:?Job ID required}"

echo "=== Job Detail: $JOB ==="
nomad_api GET "job/$JOB" | jq '{
    id: .ID,
    type: .Type,
    status: .Status,
    namespace: .Namespace,
    datacenters: .Datacenters,
    task_groups: [.TaskGroups[] | {name: .Name, count: .Count, tasks: [.Tasks[].Name]}]
}'

echo ""
echo "=== Allocations ==="
nomad_api GET "job/$JOB/allocations" | jq -r '
    sort_by(.CreateIndex) | reverse | .[:15][] |
    "\(.ID[0:8])\t\(.TaskGroup)\t\(.ClientStatus)\t\(.NodeID[0:8])\t\(.DesiredStatus)"
' | column -t

echo ""
echo "=== Failed Allocation Details ==="
for alloc_id in $(nomad_api GET "job/$JOB/allocations" | jq -r '.[] | select(.ClientStatus == "failed") | .ID' | head -3); do
    echo "--- Alloc: ${alloc_id:0:8} ---"
    nomad_api GET "allocation/$alloc_id" | jq '{
        task_states: (.TaskStates | to_entries[] | {task: .key, state: .value.State, failed: .value.Failed, events: [.value.Events[-3:][] | {type: .Type, message: .DisplayMessage}]})
    }'
done

Deployment Health

#!/bin/bash
JOB="${1:?Job ID required}"

echo "=== Latest Deployment: $JOB ==="
nomad_api GET "job/$JOB/deployment" | jq '{
    id: .ID[0:8],
    status: .Status,
    description: .StatusDescription,
    task_groups: (.TaskGroups | to_entries[] | {
        group: .key,
        desired: .value.DesiredTotal,
        placed: .value.PlacedAllocs,
        healthy: .value.HealthyAllocs,
        unhealthy: .value.UnhealthyAllocs,
        auto_revert: .value.AutoRevert
    })
}'

echo ""
echo "=== Deployment History ==="
nomad_api GET "job/$JOB/deployments" | jq -r '
    sort_by(.CreateIndex) | reverse | .[:5][] |
    "\(.ID[0:8])\t\(.Status)\t\(.StatusDescription[0:50])"
' | column -t

Client Node Status

#!/bin/bash
echo "=== Node Resource Summary ==="
for node_id in $(nomad_api GET "nodes" | jq -r '.[].ID' | head -10); do
    nomad_api GET "node/$node_id" | jq '{
        name: .Name,
        status: .Status,
        drain: .Drain,
        eligible: .SchedulingEligibility,
        datacenter: .Datacenter,
        cpu_total: .Attributes["cpu.totalcompute"],
        memory_mb: (.Resources.MemoryMB // 0),
        allocs_running: (.Attributes["nomad.allocs.running"] // "N/A")
    }'
done

echo ""
echo "=== Draining Nodes ==="
nomad_api GET "nodes" | jq -r '.[] | select(.Drain == true) | "\(.Name)\t\(.ID[0:8])\t\(.StatusDescription)"' | column -t

Namespace & ACL Overview

#!/bin/bash
echo "=== Namespaces ==="
nomad_api GET "namespaces" | jq -r '.[] | "\(.Name)\t\(.Description[0:50])"' | column -t

echo ""
echo "=== Jobs by Namespace ==="
for ns in $(nomad_api GET "namespaces" | jq -r '.[].Name'); do
    COUNT=$(nomad_api GET "jobs?namespace=$ns" | jq 'length')
    echo "$ns: $COUNT jobs"
done

echo ""
echo "=== Resource Quotas ==="
nomad_api GET "quotas" | jq -r '.[]? | "\(.Name)\t\(.Description[0:40])"' | column -t 2>/dev/null || echo "No quotas configured"

Safety Rules

• Read-only by default: Use GET requests for job, allocation, node inspection
• Never stop or purge jobs without explicit user confirmation
• Drain caution: Node drain migrates all allocations -- confirm before draining
• Never expose Nomad ACL tokens or Vault tokens in output

Output Format

Present results as a structured report:

Managing Nomad Report
═════════════════════
Resources discovered: [count]

Resource       Status    Key Metric    Issues
──────────────────────────────────────────────
[name]         [ok/warn] [value]       [findings]

Summary: [total] resources | [ok] healthy | [warn] warnings | [crit] critical
Action Items: [list of prioritized findings]

Target ≤50 lines of output. Use tables for multi-resource comparisons.

Anti-Hallucination Rules

1. NEVER assume resource names — always discover via CLI/API in Phase 1 before referencing in Phase 2.
2. NEVER fabricate metric names or dimensions — verify against the service documentation or --help output.
3. NEVER mix CLI commands between service versions — confirm which version/API you are targeting.
4. ALWAYS use the discovery → verify → analyze chain — every resource referenced must have been discovered first.
5. ALWAYS handle empty results gracefully — an empty response is valid data, not an error to retry.

Counter-Rationalizations

Shortcut	Counter	Why
"I'll skip discovery and check known resources"	Always run Phase 1 discovery first	Resource names change, new resources appear — assumed names cause errors
"The user only asked for a quick check"	Follow the full discovery → analysis flow	Quick checks miss critical issues; structured analysis catches silent failures
"Default configuration is probably fine"	Audit configuration explicitly	Defaults often leave logging, security, and optimization features disabled
"Metrics aren't needed for this"	Always check relevant metrics when available	API/CLI responses show current state; metrics reveal trends and intermittent issues
"I don't have access to that"	Try the command and report the actual error	Assumed permission failures prevent useful investigation; actual errors are informative

Common Pitfalls

• Allocation pending: Usually means insufficient resources -- check node capacity
• Failed deployments: Auto-revert may roll back silently -- check deployment history
• Stale reads: Nomad uses Raft consensus -- stale queries may return old data
• Namespace scoping: Jobs are namespace-scoped -- always specify namespace for multi-tenant clusters
• Task driver missing: Allocation fails if the required driver (docker, exec, java) is not installed on the client