/var/lib/docker/containers/xxx/yyydb-json.log is 10 GB

[ljhardy]

It looks like this log has been growing since Pulse was first started. How does / should it be cleaned up?

10G Dec 10 09:12 /var/lib/docker/containers/fa56ff528e2ae6f78abf1c0beb8747a8a678f08bcc62e9f01c749fe7609700db/fa56ff528e2ae6f78abf1c0beb8747a8a678f08bcc62e9f01c749fe7609700db-json.log

[rcourtman]

This is Docker's default logging behavior, not specific to Pulse. The json-file logging driver has no size limit by default.

To fix:

1. Add log rotation to your Docker daemon - edit /etc/docker/daemon.json:

{
  "log-driver": "json-file",
  "log-opts": {
    "max-size": "10m",
    "max-file": "3"
  }
}

Then restart Docker: systemctl restart docker

2. Or per-container in your compose file:

services:
  pulse:
    logging:
      driver: json-file
      options:
        max-size: 10m
        max-file: "3"

To clean up the existing log (stops the container temporarily):

docker compose down
truncate -s 0 /var/lib/docker/containers/fa56ff...00db/fa56ff...-json.log
docker compose up -d

Or without stopping (less safe):

truncate -s 0 /var/lib/docker/containers/fa56ff...00db/fa56ff...-json.log

[ljhardy]

Ok, will do. If this is Docker default behavior, not sure why I haven't run into this with my other containers. I will truncate and add log rotation. Thanks.

[ljhardy]

That log file sure grows fast, it's 3M in 2 hours.

[rcourtman]

3MB in 2 hours is definitely more than expected for normal operation. Pulse's default log level is info, which should not be that verbose.

To diagnose what is generating so much output:

# See what is being logged
docker logs pulse --tail 100

# Count log entries by type
docker logs pulse 2>&1 | grep -o "level.*[a-z]*" | head -50

Common causes of excessive logging:

1. Polling errors (e.g., connection refused, SSL issues) that log on every retry
2. Debug mode accidentally enabled (LOG_LEVEL=debug)
3. High poll frequency with many nodes/containers

To reduce logging:

Set LOG_LEVEL=warn in your compose file to only log warnings and errors:

services:
  pulse:
    environment:
      - LOG_LEVEL=warn

If you share a snippet of what is in the logs (sanitized), I can identify if there is a specific issue causing the verbosity.