You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Sorry for this being rather a question than an issue… the Docker watcher requires a port for each service if I understand the code right. Wouldn't it make sense to have the watcher just look for an image-tag combination and figure out the port (or ports if there are several) by itself? That way one could spin up a new deploy with a "staging" tag, test drive it, then add a "production" tag and take down the old deploy. Or am I missing something?
The text was updated successfully, but these errors were encountered:
Presumably so if you have a single container with multiple open ports(say, an http server and some type of monitoring daemon) you can discover them both and apply different backend settings in haproxy. Hence, service discovery rather than container discovery.
Concerning the image-tag combination -- I do currently deploy this way and have #96 open for the feature. This also provides the benefit of preventing conflicting image versions of an application from being used as valid backends if your orchestration tool or sysadmin doesn't always clean up after itself.
Sorry for this being rather a question than an issue… the Docker watcher requires a port for each service if I understand the code right. Wouldn't it make sense to have the watcher just look for an image-tag combination and figure out the port (or ports if there are several) by itself? That way one could spin up a new deploy with a "staging" tag, test drive it, then add a "production" tag and take down the old deploy. Or am I missing something?
The text was updated successfully, but these errors were encountered: