WebServer New Added 26 September 2026

nginx: 1024 worker_connections are not enough

Each worker has a fixed number of connection slots and a proxied request uses two of them, one to the client and one to the upstream, so the ceiling arrives at roughly half the concurrency you expected. New connections queue or fail while CPU and memory look fine, and the message repeats once per worker in the error log.

Quick fix

Read the commands before running them. Anything that restarts a service, deletes data or changes permissions should be tried on a non-production system first.

Quick fix

# Measure the ceiling rather than guessing at it
curl -s localhost/nginx_status      # requires stub_status
ss -tan state established | wc -l

# events block, not http: proxied requests take two slots each
events {
    worker_connections 8192;
}
worker_processes auto;
worker_rlimit_nofile 32768;         # must exceed worker_connections

# The file descriptor limit is the real cap, and systemd sets its own
systemctl show nginx -p LimitNOFILE
# override with a drop in: [Service] LimitNOFILE=65535

sudo nginx -t && sudo systemctl reload nginx

# Keepalive to upstreams removes most of the second slot's churn
# upstream app { server 127.0.0.1:3000; keepalive 64; }

How to diagnose WebServer errors

Web server errors are usually about permissions, binding, or configuration that was never loaded. A common trap: the server runs as an unprivileged user (www-data, nginx), so it needs execute permission on every directory in the path to a file, not just read permission on the file itself. Another: editing a config file changes nothing until the server is reloaded and the file is actually included.

If the quick fix above does not resolve it, work through these steps. They apply to this whole class of error, not just to this one message, which is usually what saves the time.

Validate configuration before reloading: nginx -t, apachectl configtest, caddy validate. This catches syntax errors without dropping traffic.
Dump the fully resolved configuration with nginx -T to see what is actually in effect, including every include.
Read the error log, not the access log, for 5xx causes. nginx names the failing upstream and the exact filesystem path it could not open.
For permission errors, test as the server user: sudo -u www-data cat /path/to/file. Check execute bits on every parent directory.
For binding failures, find the current listener with ss -tulpn | grep :80. Ports below 1024 need privileges or a capability such as CAP_NET_BIND_SERVICE.

Tools worth reaching for

nginx -t / nginx -T
apachectl configtest
ss -tulpn
tail -f error.log
sudo -u www-data

Authoritative references

Primary documentation for this error, worth reading before applying any fix in production.

nginx.org nginx.org

Related WebServer errors

See all 20 WebServer errors →

Browse other categories

Something missing or wrong?

This entry is maintained by hand. If the fix is out of date, incomplete, or you have a better one, email a correction and it will be reviewed.