Skip to content

Fleet Management

Orchestrate commands across multiple VMs with parallel execution and intelligent routing.

Overview

Fleet commands enable distributed operations across VM fleets with pattern matching, parallel execution, and comprehensive error handling.

Available Commands

Distributed Operations

Quick Start Examples

Execute on All VMs

# Run command on every VM
azlin batch exec "*" "docker ps"

# Check disk space across fleet
azlin batch exec "*" "df -h"

# Update all VMs
azlin batch exec "*" "sudo apt update && sudo apt upgrade -y"

Pattern Matching

# Target specific VM groups
azlin batch exec "api-*" "systemctl restart nginx"
azlin batch exec "db-*" "docker logs postgres"
azlin batch exec "test-*" "rm -rf /tmp/*"

# Multiple patterns
azlin batch exec "web-*,api-*" "git pull origin main"

Specific VM Lists

# Comma-separated VM names
azlin batch exec "vm1,vm2,vm3" "uptime"

# Tag-based selection
azlin batch exec --tag env=prod "systemctl status myapp"

Monitoring Fleet

# Who's logged in across fleet
azlin w

# All running processes
azlin ps

# Filter by pattern
azlin w --filter "api-*"
azlin ps --filter "db-*"

Key Features

Parallel Execution

Execute commands simultaneously across VMs:

# Parallel updates
azlin batch exec "*" "git pull" --parallel 10

# Sequential for safety
azlin batch exec "*" "systemctl restart app" --sequential

Error Handling

Continue execution even if some VMs fail:

# Continue on errors
azlin batch exec "*" "risky-command" --continue-on-error

# Stop on first failure
azlin batch exec "*" "critical-command" --stop-on-error

Output Aggregation

Collect and format results:

# Show all output
azlin batch exec "api-*" "docker ps" --verbose

# Summary only
azlin batch exec "*" "uptime" --summary

# Save to file
azlin batch exec "*" "systemctl status" > fleet-status.txt

File Distribution

Sync files to multiple VMs:

# Sync to all VMs
azlin batch sync "*" ~/myproject

# Sync to specific pattern
azlin batch sync "web-*" ~/frontend --delete

# Dry-run first
azlin batch sync "*" ~/config --dry-run

Common Workflows

Deploy Application Update

# 1. Sync code to all app servers
azlin batch sync "app-*" ~/myapp

# 2. Restart services
azlin batch exec "app-*" "systemctl restart myapp"

# 3. Verify status
azlin batch exec "app-*" "systemctl status myapp"

Fleet Health Check

# Check system resources
azlin batch exec "*" "df -h && free -h && uptime"

# Check application status
azlin batch exec "*" "systemctl status myapp"

# Check docker containers
azlin batch exec "*" "docker ps && docker stats --no-stream"

Configuration Management

# Update configuration
azlin batch sync "*" ~/config/nginx.conf /etc/nginx/

# Reload services
azlin batch exec "*" "sudo systemctl reload nginx"

# Verify configuration
azlin batch exec "*" "nginx -t"

Log Collection

# Collect application logs
azlin batch exec "*" "tail -n 100 /var/log/myapp.log" > all-logs.txt

# Search for errors
azlin batch exec "*" "grep ERROR /var/log/myapp.log"

# Download logs
for vm in $(azlin list --name-only); do
  azlin connect $vm --command "cat /var/log/myapp.log" > logs/${vm}.log
done

Pattern Matching Reference

Wildcard Patterns

# All VMs starting with "api-"
azlin batch exec "api-*" "command"

# All VMs ending with "-prod"
azlin batch exec "*-prod" "command"

# All VMs containing "test"
azlin batch exec "*test*" "command"

Multiple Patterns

# Match multiple patterns (OR)
azlin batch exec "web-*,api-*,db-*" "command"

# Exclude pattern (requires explicit list)
# List all VMs, then filter manually
azlin list --name-only | grep -v "test-" | xargs -I {} azlin connect {} --command "..."

Tag-Based Selection

# Select by tag
azlin batch exec "*" "command" --tag env=prod

# Multiple tags
azlin batch exec "*" "command" --tag env=prod --tag app=web

Performance Tuning

Parallelism

# Default: 5 parallel executions
azlin batch exec "*" "command"

# More parallel (faster, more load)
azlin batch exec "*" "command" --parallel 20

# Sequential (slower, more reliable)
azlin batch exec "*" "command" --sequential

Timeout Control

# Set timeout per VM
azlin batch exec "*" "long-command" --timeout 300

# No timeout (wait forever)
azlin batch exec "*" "command" --no-timeout

Retry Logic

# Retry failed commands
azlin batch exec "*" "flaky-command" --retry 3

# Retry delay
azlin batch exec "*" "command" --retry 3 --retry-delay 5

Best Practices

1. Test with Dry-Run

# Always test sync operations
azlin batch sync "*" ~/important-data --dry-run
azlin batch sync "*" ~/important-data

2. Use Patterns Wisely

# Good: Specific patterns
azlin batch exec "api-prod-*" "restart-command"

# Risky: Too broad
azlin batch exec "*" "dangerous-command"

3. Monitor Output

# Use verbose for important operations
azlin batch exec "*" "critical-command" --verbose

# Save output for auditing
azlin batch exec "*" "command" | tee operation.log

4. Handle Failures

# Critical operations: Stop on error
azlin batch exec "*" "database-migration" --stop-on-error

# Best-effort operations: Continue on error
azlin batch exec "*" "optional-update" --continue-on-error

Fleet Monitoring Commands

System Status

# Logged-in users
azlin w

# All processes
azlin ps

# System resources
azlin batch exec "*" "top -bn1 | head -20"

Application Status

# Service status
azlin batch exec "*" "systemctl status myapp"

# Docker containers
azlin batch exec "*" "docker ps"

# Application health
azlin batch exec "*" "curl -sf http://localhost:8080/health"

Logs and Diagnostics

# Recent logs
azlin batch exec "*" "journalctl -n 50"

# Error logs
azlin batch exec "*" "grep ERROR /var/log/myapp.log | tail -20"

# Disk usage
azlin batch exec "*" "df -h"

Error Handling

Common Issues

Connection failures: 

# Some VMs unreachable
azlin batch exec "*" "command" --continue-on-error

Command failures: 

# Track which VMs failed
azlin batch exec "*" "command" --verbose 2>&1 | grep ERROR

Timeout issues: 

# Increase timeout
azlin batch exec "*" "slow-command" --timeout 600

See Also