Managing Fly Io Deep

Fly.io Deep Management

Comprehensive analysis of Fly.io apps, machines, volumes, and global deployment status.

Phase 1: Discovery

#!/bin/bash
TOKEN="${FLY_API_TOKEN}"
BASE="https://api.machines.dev/v1"
AUTH=(-H "Authorization: Bearer ${TOKEN}" -H "Content-Type: application/json")

echo "=== Apps Inventory ==="
flyctl apps list --json 2>/dev/null \
  | jq -r '.[] | "\(.Name)\t\(.Organization.Slug)\t\(.Status)\t\(.Deployed)\t\(.Hostname)"' \
  | column -t | head -20

echo ""
echo "=== Machines per App ==="
for APP in $(flyctl apps list --json 2>/dev/null | jq -r '.[].Name'); do
  curl -s "${BASE}/apps/${APP}/machines" "${AUTH[@]}" \
    | jq -r ".[] | \"${APP}\t\(.id)\t\(.state)\t\(.region)\t\(.config.guest.cpus)cpu/\(.config.guest.memory_mb)MB\t\(.config.image | split(\":\") | last)\"" 2>/dev/null
done | column -t | head -30

echo ""
echo "=== Volumes ==="
for APP in $(flyctl apps list --json 2>/dev/null | jq -r '.[].Name'); do
  flyctl volumes list --app "$APP" --json 2>/dev/null \
    | jq -r ".[] | \"${APP}\t\(.id)\t\(.name)\t\(.region)\t\(.size_gb)GB\t\(.state)\"" 2>/dev/null
done | column -t | head -20

echo ""
echo "=== Certificates ==="
for APP in $(flyctl apps list --json 2>/dev/null | jq -r '.[].Name'); do
  flyctl certs list --app "$APP" --json 2>/dev/null \
    | jq -r ".[] | \"${APP}\t\(.hostname)\t\(.clientStatus)\t\(.source)\"" 2>/dev/null
done | column -t | head -20

Phase 2: Analysis

#!/bin/bash
TOKEN="${FLY_API_TOKEN}"
BASE="https://api.machines.dev/v1"
AUTH=(-H "Authorization: Bearer ${TOKEN}" -H "Content-Type: application/json")

echo "=== Machine Health Checks ==="
for APP in $(flyctl apps list --json 2>/dev/null | jq -r '.[].Name'); do
  for MACHINE in $(curl -s "${BASE}/apps/${APP}/machines" "${AUTH[@]}" | jq -r '.[].id' 2>/dev/null); do
    curl -s "${BASE}/apps/${APP}/machines/${MACHINE}" "${AUTH[@]}" \
      | jq "{app: \"${APP}\", machine: .id, state: .state, checks: [.checks[]? | {name, status, output: .output[0:50]}]}" 2>/dev/null
  done
done | head -30

echo ""
echo "=== Region Distribution ==="
for APP in $(flyctl apps list --json 2>/dev/null | jq -r '.[].Name'); do
  echo "--- ${APP} ---"
  curl -s "${BASE}/apps/${APP}/machines" "${AUTH[@]}" \
    | jq -r '.[].region' 2>/dev/null | sort | uniq -c | sort -rn
done

echo ""
echo "=== Autoscaling Config ==="
for APP in $(flyctl apps list --json 2>/dev/null | jq -r '.[].Name'); do
  flyctl autoscale show --app "$APP" --json 2>/dev/null \
    | jq "{app: \"${APP}\", min: .MinCount, max: .MaxCount, balanceRegions: .BalanceRegions}" 2>/dev/null
done

echo ""
echo "=== Secrets (names only) ==="
for APP in $(flyctl apps list --json 2>/dev/null | jq -r '.[].Name'); do
  SECRETS=$(flyctl secrets list --app "$APP" --json 2>/dev/null | jq -r '.[].Name' 2>/dev/null | tr '\n' ', ')
  [ -n "$SECRETS" ] && echo "${APP}: ${SECRETS}"
done

echo ""
echo "=== Resource Summary ==="
for APP in $(flyctl apps list --json 2>/dev/null | jq -r '.[].Name'); do
  curl -s "${BASE}/apps/${APP}/machines" "${AUTH[@]}" \
    | jq "{app: \"${APP}\", machines: length, total_cpus: [.[].config.guest.cpus] | add, total_memory_mb: [.[].config.guest.memory_mb] | add}" 2>/dev/null
done

Output Format

FLY.IO DEEP ANALYSIS
======================
App              Machines  Regions    CPU/Mem       Volumes  Status    Checks
──────────────────────────────────────────────────────────────────────────────
web-api          3         iad,lhr    2cpu/512MB    0        deployed  passing
worker           2         iad        4cpu/1024MB   2x10GB   deployed  passing
cron-job         1         iad        1cpu/256MB    0        deployed  passing

Regions: iad(4) lhr(2) | Total: 6 machines, 10 CPUs, 2.5GB RAM
Certificates: 3 valid | Secrets: 12 across 3 apps

Safety Rules

• Read-only: Only use flyctl * list, show and GET endpoints
• Never deploy, scale, or destroy machines without explicit confirmation
• Secrets: Never output secret values, only list names
• Rate limits: Fly Machines API has no published rate limit but use reasonable request rates

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