enable-long-lived-container-logs.md

Enable Long-Lived Container Logs

An all-too-common, frustrating and time-consuming, pain point: tailed container logs terminate silently in a seemingly non-deterministic manner, forcing you to terminate your connection to the container and execute the logs command again. A workaround to prevent this behavior (when using docker-compose) is to use the COMPOSE_HTTP_TIMEOUT environment variable¹, and increase the default timeout value to a more suitable duration, in seconds:

COMPOSE_HTTP_TIMEOUT=86400 docker-compose logs -ft --tail="all" --no-log-prefix $CONTAINER_NAME

Footnotes

1. COMPOSE_HTTP_TIMEOUT was deprecated in docker-compose v2. I'm unsure if this workaround still works in versions >= 2, or what an equivalent workaround would be. ↩