Load Average is a number showing demand on a Linux server's CPU and other resources. It's the key performance metric. Shown by the uptime or top commands: "load average: 0.50, 0.75, 1.20" β 1, 5, and 15-minute averages.
How to interpret Load Average
Meaning depends on the number of CPU cores. On a 1-core server: LA 1.0 β CPU is 100% busy; 2.0 β demand is 2x more (50% queued); 0.5 β 50% busy, 50% idle.
On a 4-core server: LA 4.0 β 100% busy; 8.0 β 2x demand; 2.0 β 50% busy.
Rule: Load Average / number of cores = CPU utilization percentage. Example: on an 8-core server LA = 4.0 β 50% load.
When to worry
High Load Average (above core count) slows sites down. Warning signs: 1-minute LA > cores Γ 1.5 β temporary spike; 5-minute LA > cores Γ 1 β sustained load problem; 15-minute LA > cores β long-term issue.
Causes of high load
WordPress brute force attack β bots send many requests to wp-login.php. Fix: IP restrictions on wp-login or Wordfence.
MySQL slow queries β badly indexed database. Fix: enable slow_query_log and analyze with EXPLAIN.
New bot crawls β Google, Yandex, and other bots crawling the site simultaneously. Fix: robots.txt crawl-delay or Cloudflare rate limiting.
Cron job collisions β many heavy crons running at the same time.
Monitoring tools
uptime β simplest: shows 1, 5, 15-minute LA.
top β real-time view of CPU, RAM, and processes.
htop β an improved top (covered in a separate article).
Munin, Nagios, Prometheus + Grafana β professional monitoring for large servers.
Sayt.uz practice
Sayt.uz servers run 24/7 monitoring: an automatic alert fires when Load Average exceeds the 80% threshold. Shared hosting averages 0.5-2.0 LA (normal for 4-core servers). On high load we contact the client and find the cause. Tip: if you run VPS or a dedicated server, daily monitoring is essential β load can spike suddenly.