Didn't we conclude in a recent retro that we always want to deploy updates images to staging first, before production? Did we make an exception here because Tunnelbroker is barely live in production?

In D12953#366484, @ashoat wrote:

Didn't we conclude in a recent retro that we always want to deploy updates images to staging first, before production? Did we make an exception here because Tunnelbroker is barely live in production?

Most of the logic I deployed on prod (notifs work) was on the staging for some time and was tested by my and @marcin,
For the invalid device token work, I deployed this on staging and spent most of the day testing this before releasing, additionally, this is notifs- related error handling code, it's an additional feature and is not affecting existing logic which is live, and for some time it won't be executed so decided to deploy both.

But overall, I agree with you - I should've bump staging to 0.14 and prod to 0.12 (not 0.14) to keep our deploying process consent, thanks for pointing that out.