[ENHANCE] ArgoCD "restart" instead of starting a new rollout for blue/green.

[im-aIex]

Is your feature request related to a problem? Please describe.
When using long-lived / manual blue/green rollouts it currently creates a new green rollout and requires time / manual intervention to make it to the blue pods. Having the option to run a "restart" would allow all pods in an existing blue/green rollout to be recreated and still maintain HA.

Describe the solution you'd like
Add an option (annotation/strategy) to run a "restart" instead of just causing an update that forces a new rollout.

Additional context
See Restarting Rollout Pods for more details.

Both of these mechanisms updates the Rollout's .spec.restartAt to the current time in the form of a RFC 3339 formatted UTC string (e.g. 2020-03-30T21:19:35Z), which indicates to the Rollout controller that all of a Rollout's Pods should have been created after this timestamp.

[jangeja]

Same applies for canary rollouts.

If we have a long running canary, and we update a secret, only the canary gets updated.

This can present itself a major problem if the stable RS is not reloaded with the updated secrets/CMs.

Can probably fix this with a simple change to the updateRollout Callback

Ill push a PR shortly

[jangeja]

This can be closed, once a release is cut witht he merged PR.

[jangeja]

@MuneebAijaz whats the release cadence for reloader ? Would love to close this issue out!

[MuneebAijaz]

@jangeja images tagged as merge-728 https://github.com/stakater/Reloader/pkgs/container/reloader should have this change for immediate use. as for the release, i was hoping to resolve charts release workflow as well before making a manual release but could be done earlier, let's see.

[jangeja]

Thank you! Will use the merge-728 image until a release is made.

[im-aIex]

Fix released in v1.1.0