Skip to content

Releases: jittering/traefik-kop

v0.14: Namespaces

11 Jul 13:16
Compare
Choose a tag to compare

Changelog

New Features

  • Allow filtering containers processed by traefik-kop using namespaces

Fixes

  • Use exact service name match when searching container labels (#39, thanks @damfleu)

Full Changelog: v0.13.3...v0.14

v0.13.3

30 Jun 22:31
Compare
Choose a tag to compare

Changelog

  • 16beda8 build: bump go version to 1.22

v0.13.2

09 Oct 15:25
Compare
Choose a tag to compare

Changelog

  • 10ab916 fix: properly stringify floats when writing to redis (resolves #25)

v0.13.1

v0.13

17 Oct 00:50
437a95d
Compare
Choose a tag to compare

Changelog

Full Changelog

New Features

  • Set bind IP per-container or service
  • Set traefik docker provider config (e.g., defaultRule)

Fixes

  • Correctly set port for TCP and UDP services

Closed issues

  • Go runtime error #20
  • Default Rule #18
  • Provide IP for each docker via label #17
  • setting port for tcp service does not work #16
  • Doesn't work with multiple services on one container #14

v0.12.1

10 Sep 19:18
Compare
Choose a tag to compare

Changelog

This release updates the upstream version of the traefik library to v2.8.4 and adds additional logging around port detection (both debug and info levels) to make it easier to see what's going on and troubleshoot various scenarios.

Commits

  • 8c5a3f0 build: bump actions/cache to v3
  • dad6e90 build: bump go version in github actions
  • f009b84 docs: added more detail and logging around port selection
  • 2f18114 test: added helloworld service for testing multiple bindings
  • be636f7 build: upgraded traefik to 2.8.4 (now supports go 1.18+)

v0.12

22 Jun 17:40
Compare
Choose a tag to compare

Changelog

Notes

By default, traefik-kop will listen for push events via the Docker API in
order to detect configuration changes. In some circumstances, a change may not
be pushed correctly. For example, when using healthchecks in certain
configurations, the start -> healthy change may not be detected via push
event. As a failsafe, there is an additional polling mechanism to detect those
missed changes.

The default interval of 60 seconds should be light so as not to cause any
issues, however it can be adjusted as needed via the KOP_POLL_INTERVAL env var
or set to 0 to disable it completely.

Full Changelog

  • 347352b build: fix goreleaser tidy
  • b6447c3 build: go mod tidy
  • 12ad255 docs: added poll interval to readme
  • 10f7aab feat: expose providers in case anyone wants to reuse
  • 5b58547 feat: add log message when explicitly disabling polling
  • 02802d5 feat: configurable poll interval (default 60)
  • b2ef52b feat: combine providers into single config watcher
  • 07fe8aa feat: added polling provider as a workaround for healthcheck issue
  • cc3854b feat: added config for changing docker endpoint
  • c309d40 build: upgraded traefik lib to v2.7
  • 32c2df6 test: added pihole container (with builtin healthcheck)
  • e770242 docs: updated changelog

v0.11

17 Jun 15:36
Compare
Choose a tag to compare

Changelog

Full Changelog

Notes

  • If your container is configured to use a network-routable IP address via an
    overlay network or CNI plugin, that address will override the bind-ip
    configuration when the traefik.docker.network label is present.

Merged pull requests:

  • Add support for traefik.docker.network #8 (hcooper)

v0.10.1

12 Mar 23:54
e0af6eb
Compare
Choose a tag to compare

Changelog

  • e0af6eb Merge pull request #7 from jittering/fix/port-detect

v0.10.0

11 Feb 01:18
Compare
Choose a tag to compare

Changelog

  • 5d029d2 feat: add support for ports published via --publish-all (closes #6)