I am using a fileprovider with one big dynamic config file for Traefik. That file is built into the image. I need to change the file regularly on a weekly basis, adding new entries. This is done by building a new image, and replacing the existing Traefik instance (which runs in a Fargate cluster) to a new one gracefully. If traefik has a problem, and does not respond to heartbeat, then the rollout just stops gracefully. However, when I misstype one space in the big dynamic conf yaml, the new instance starts happily, and no routing willhappen at all. Is there a way to make the Traefik fail the healthcheck on such occasion? My other option is to read the config file from an external place, where a runtime change is possible. I have done some manual testing, not using a fargate cluster, and Traefik is intelligent enough to just ignore the invalid dynamic yaml change and keep working with the old settings. I have also split the config into smaller files,so that if one file is wrong, the other routings can continue working. This works well on runtime change.However, when I restart the traefik, and one out of the many dynamic configs is wrong, the whole traefik won't work. So this solution is also not good. Do you have any suggestions what else I could do?
Either make the healthcheck fail or make Traefik start by ignoring only the invalid configs.
Thank you