Change in atlassian-connect.json baseUrl property is not reflected in user instances


We recently changed the BaseURL of the atlassian-connect.json of our Git Flow Chart app for Bitbucket cloud. We recreated the app and notified users that they should reinstall it in their instance to get the latest & greatest. However, we have been receiving support requests that this is not working: they are still seeing the old app, even after re-installing it. It seems like the changes to the baseUrl in atlassian-connect.json are not being picked up by the Marketplacebot, or at least it is not resulting in an update of the user environment.

I’ve raised, please help!

Hi Remie,
The best way to get help on this sort of thing is to use the service desk for which you posted your ticket. Posting here in addition to the ticket just creates more work for us and doesn’t really offer any help to the community.

Also, the team that can make changes for this works on Pacific time, so we won’t really have a chance to fix much this early.


@ehenry you make valid points sir. And I really wish I would not have to raise this issue here. But there are a few reasons why I did it anyway.

  1. There is a documented lack of response on DEVHELP tickets lately, with many vendors feeling like there is no way for them to reach Atlassian, especially on tickets that directly impact customers.

  2. The DEVHELP tickets are private, meaning that there is no way for other vendors to indicate whether or not they experience the same issue. There is no option for me to create a public ticket for them to vote on, so this is the only way I can signal to other vendors about this issue and have them take notice / share their experiences. I know for a fact that others have also experienced this.

Also, please note that it is not possible for me to know which team does what in which timezone. I would assume that DEVHELP tickets are vetted by others than the bitbucket team. I would also assume that a P1 issue does not sit for 9 hours without any response at all.

I would love to not have to use a multi-channel approach. But I do not feel I have the tools to actually reach Atlassian in a proper manner.


@remie Thanks for bringing this up. I’ll see what I can do to help improve this.

If you are having trouble with responses being slow, community is definitely the place to check to see if others are experiencing the same thing.

I understand that you wouldn’t have any knowledge about the timezone for this sort of thing. Unfortunately, Bitbucket cloud is in a unique place for this because the Atlassian Marketplace is not synced with the Bitbucket cloud marketplace (we’re working on this). I also don’t know what the SLA is supposed to be for DEVHELP tickets, but I can follow up on it.

@remie - as @ehenry mentioned, things are not in sync, and that’s currently a WIP. Meanwhile, it should be resolved now for your app. I’ll follow up with the Bitbucket Cloud team to see what’s possible for mitigation until the final fix is in place.

Re: escalations – at this time, only app security issues have an escalation path/workflow. And by the way, I don’t suggest using that as a ticket type as a workaround. :slight_smile: Right now, the team that is responsible for DEVHELP bugs/issues is based in APAC (you know @acalantog and @iragudo). We’ll discuss strategies on improving escalation of issues like this one.


@nmansilla we have migrated our Jira Cloud app to a new base URL more than one month ago, yet we still see a handful of regular requests to the old URL. It seems those hosts did not get migrated automatically to the new base URL. Is there any manual action required by Atlassian to fix this? How should we proceed with this?

Just following up on this. @dmorrow sorry to tag you directly but I am not sure who else to contact to get some movement on this. Would you mind forwarding this to the right people within Atlassian?

Hi @tbinna,

Is the app listed in Marketplace and does it appear that Marketplace has re-installed the app in a number of tenants, but not all? Have there been any subsequent changes to the app that Marketplace should have triggered a re-install for? Could it be that certain tenants have not been upgraded because the installed version of your app pre-dates some changes that requires admin approval such as an increase of scopes or a change to the payment model? Changing the version attribute in the app’s Connect descriptor will force Marketplace into another round of installation attempts and may be worthwhile if we believe the reason for the failed installations is transient.


1 Like