Network Automation Using Python

SSH-based toolkit for routine Cisco IOS administration — health checks, interface validation, connectivity tests, config backups, and simple health reports.

Developed against a small lab of Catalyst switches and an ISR router. Works with Netmiko for day-to-day sessions, with a Paramiko fallback when Netmiko can't negotiate a session cleanly.

---

Project overview

Manual checks across a handful of switches get old fast. This project automates the boring parts:

• Pull hostname, IOS version, uptime, CPU, memory, serial, and model
• Flag interfaces that are admin-up / protocol-down, or accumulating CRC / input errors
• Ping gateway, DNS, and a management server from each device
• Archive running-config and startup-config
• Emit CSV + JSON + text summaries so results are easy to skim or hand to someone else

It's intentionally small. No inventory database, no web UI — YAML in, reports out.

---

Architecture

devices.yaml ──► inventory.py ──► main.py (CLI)
                                      │
                                      ▼
                               ssh_manager.py
                           (Netmiko → Paramiko)
                                      │
          ┌─────────────┬─────────────┼─────────────┬──────────────┐
          ▼             ▼             ▼             ▼              ▼
   health_check  interface_val  connectivity  config_backup  report_generator

Module	Role
ssh_manager.py	Session lifecycle, enable mode, Paramiko fallback
health_check.py	show version / CPU / memory parsing
interface_validator.py	show interfaces + issue rules
connectivity.py	Device-originated pings
config_backup.py	Config archival under backups/
report_generator.py	CSV / JSON / text output
demo_data.py	Canned CLI output for offline demos

---

Folder structure

network-automation/
├── README.md
├── requirements.txt
├── .gitignore
├── config/
│   ├── devices.yaml      # inventory
│   └── settings.yaml     # timeouts, thresholds, ping targets
├── reports/              # generated reports land here
├── logs/                 # automation.log
├── backups/              # running/startup configs
├── output/               # scratch / misc output
├── src/
│   ├── main.py
│   ├── ssh_manager.py
│   ├── health_check.py
│   ├── interface_validator.py
│   ├── connectivity.py
│   ├── config_backup.py
│   ├── inventory.py
│   ├── report_generator.py
│   ├── demo_data.py
│   ├── utils.py
│   └── exceptions.py
└── tests/

---

Requirements

• Python 3.9+
• Network reachability to devices on TCP/22 (or your SSH port)
• Local credentials with privilege to run show commands (enable secret if needed)

Python packages (see requirements.txt):

• netmiko
• paramiko
• PyYAML
• pytest (tests only)

---

Installation

git clone <repo-url>
cd network-automation   # or repo root
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

---

Configuring devices.yaml

Edit config/devices.yaml:

devices:
  - name: core-sw-01
    host: 192.168.10.1
    device_type: cisco_ios
    username: netadmin
    password: "LabPass123"
    secret: "LabEnable123"
    port: 22
    site: hq-core
    role: distribution

Notes:

• device_type must be a Netmiko platform key (cisco_ios for most Catalyst/ISR boxes).
• Prefer keeping real passwords out of git — copy to config/devices.local.yaml (gitignored) or swap credentials before running.
• Optional fields (site, role) are informational and show up in list-devices.

Global knobs live in config/settings.yaml (timeouts, CPU/memory thresholds, ping targets, demo_mode).

---

Example commands

# inventory sanity check
./run.sh list-devices
# or: python src/main.py list-devices

# health only (one device)
python src/main.py health -d core-sw-01

# interfaces / connectivity / backups
python src/main.py interfaces
python src/main.py connectivity
python src/main.py backup -d edge-rtr-01

# everything + reports
python src/main.py full-report

# offline demo (no SSH) — uses canned show/ping output
python src/main.py --demo full-report

# debug logging
python src/main.py -v health

Exit codes: 0 ok, 1 usage/inventory error, 2 one or more devices failed checks.

---

Example output

$ python src/main.py --demo full-report

=== Summary ===
Devices checked : 4
Failed devices  : access-sw-02
Warnings        : 1
Interface issues: 4
CSV report      : reports/health_report_20260728_143022.csv
JSON report     : reports/full_report_20260728_143022.json
Text summary    : reports/health_summary_20260728_143022.txt

Text summary excerpt:

============================================================
NETWORK HEALTH SUMMARY
============================================================
Generated: 2026-07-28T14:30:22
Devices checked: 4

Failed devices (1):
  - access-sw-02

Warnings (1):
  - access-sw-02: High CPU (5min avg 81%)

Interface issues (4):
  - access-sw-01 GigabitEthernet0/2: admin up but line protocol down
  - access-sw-01 GigabitEthernet0/5: CRC errors: 188
  - access-sw-02 GigabitEthernet0/10: CRC errors: 940
  ...

Sample reports from a lab/demo run are checked in under reports/sample/.

---

Screenshots

Placeholder — drop terminal captures here when documenting a lab run.

---

Known limitations

• Parsing is regex-based against classic IOS output. IOS-XE / NX-OS quirks will need tweaks.
• Paramiko fallback is intentionally minimal (no full Netmiko feature parity).
• No parallel device sessions yet — large inventories will be slow (fine for a small lab batch).
• Interface validation skips Null / Loopback / Vlan by default; adjust if you care about SVIs.
• Credentials in YAML are plain text. Fine for a lab, not fine for production secrets management.

---

Future improvements

• Threaded / async device workers
• Diff last two config backups and email on change
• NAPALM getters as an optional backend
• Slack / email notification hook for failed devices
• Richer interface baselines (expected speed/duplex per port)

---

Tests

pytest -q

Parser tests run fully offline.

---

License

Personal / portfolio project. Use and adapt as you like.