Is there an easy way to reduce HTTP request logging?

[minorai]

Currently it seems that the nginx setup in the frigate container logs every request to stdout regardless if it is successful or failed. This clutters docker logs especially with health probes. Is there an option to cut down on the logs without messing with nginx configuration files inside the image?

STDOUT 2026-02-21 20:25:50.325190284  172.21.0.1 - - [21/Feb/2026:20:25:50 +0000] "GET / HTTP/1.1" 200 6835 "-" "Blackbox-Exporter/0.28.0" "-"
STDOUT 2026-02-21 20:26:05.325767718  172.21.0.1 - - [21/Feb/2026:20:26:05 +0000] "GET / HTTP/1.1" 200 6835 "-" "Blackbox-Exporter/0.28.0" "-"
STDOUT 2026-02-21 20:26:20.325958865  172.21.0.1 - - [21/Feb/2026:20:26:20 +0000] "GET / HTTP/1.1" 200 6835 "-" "Blackbox-Exporter/0.28.0" "-"
STDOUT 2026-02-21 20:26:35.325968536  172.21.0.1 - - [21/Feb/2026:20:26:35 +0000] "GET / HTTP/1.1" 200 6835 "-" "Blackbox-Exporter/0.28.0" "-"
STDOUT 2026-02-21 20:26:50.325721295  172.21.0.1 - - [21/Feb/2026:20:26:50 +0000] "GET / HTTP/1.1" 200 6835 "-" "Blackbox-Exporter/0.28.0" "-"

Images used is stable-tensorrt
Adding

logging:
  default: warning

did not change anything

[inkeep-ai-assistant[bot]]

The logger configuration you tried only controls Frigate's Python application logging, not nginx HTTP access logs(1). The nginx logs are separate and aren't affected by the logger.default setting(1).

To reduce nginx logging without modifying files inside the container, you can use a bind mount or volume map of Frigate's nginx config(2). You could check the user agent header and disable logging specifically for your health probe requests(2).

For example, you could create a custom nginx configuration that filters out requests from your "Blackbox-Exporter" user agent, then mount it over the default nginx config location(3). When testing nginx config changes, you can copy and reload the config with:

sudo cp docker/main/rootfs/usr/local/nginx/conf/* /usr/local/nginx/conf/ && sudo /usr/local/nginx/sbin/nginx -s reload

(3)

However, there isn't a built-in Frigate configuration option to control nginx access logging without customizing the nginx configuration files.

📚 Sources:

1. Advanced Options
2. Exclude the healthcheck endpoint from nginx logs
3. Contributing To The Main Code Base

---

Have another question? Just tag @inkeep.

[minorai]

created a workaround that creates a custom docker image by patching existing nginx.conf

FROM ghcr.io/blakeblackshear/frigate:stable-tensorrt

COPY ./script.sh /script.sh
RUN chmod +x /script.sh && /script.sh /usr/local/nginx/conf/nginx.conf && rm /script.sh

#!/bin/bash

CONFIG_FILE="$1"

if [[ ! -f "$CONFIG_FILE" ]]; then
    echo "File not found: $CONFIG_FILE"
    exit 1
fi

# Insert 'access_log off;' after the first 'http {' line
sed -i '/http\s*{/a \
    access_log off;' "$CONFIG_FILE"