Atlassian Stack moving to PostgreSQL Only?

[Marcus Herzog]

I just got back from Summit. What a nice week. I always enjoy visiting america.

Anyway, what I wanted to ask is about something that I heard  two guys that were using Atlassian badges talking. They were talking about JIRA and other apps going to support only PostgreSQL in the near future.

Is this true or just two guys dreaming?

Such a move would be a massive deal breaker for us. Our company policy only allows us to use SQL Server and such move would prohibit us from using JIRA in the future.

Please respond as soon as possible as we won't start upgrading to JIRA 7 if that is the case. We will start looking for alternatives.

[Otto]

Hi Marcus,

First up let me provide clarity: Atlassian has not made a decision to consolidate it's supported databases for JIRA Server or other Atlassian products.

At Summit you would have heard that one of Atlassian's big focus areas is improving our Cloud products so that they are highly performant, reliable, secure, capable, extensible, compliant, etc... so that all of our Customers will eventually be able to move to Atlassian Cloud products. We're going to be investing significantly in getting the products to that state. From time to time that means we need to look at the overall tech debt we carry so we can focus our development efforts to get to that cloud state quickly enough for our customers.

As you may or may not know, PostgreSQL is the DB we use to support our internal and our customer cloud instances. So this means PostgreSQL gets more dogfooding and real world usage than any of the other supported databases. Consequently, one of the discussions the JIRA team has been having recently is in regard to a cloud project where our development effort is doubled due to having to support multiple databases.

We know and appreciate that there is a lot of customer sensitivity here, as per you points on being an "MS SQL shop" and are not taking that decision lightly. That's why I and some of my colleagues were using Summit as a sounding board for how a move to PostgreSQL only support might impact customers.

To that end, we'd love to hear more from you and others about our supported DBs. Everyone can provide their input by filling out this very short survey.

[Nic Brough -Adaptavist-]

I think the survey has locked down permissions - Atlassian only? (Is there a space to suggest "JIRA wiki" in there as well?)

[Otto]

Survey issues resolved... here's the link: https://docs.google.com/a/atlassian.com/forms/d/1MndjeXU9sowSDn6v7Il5RP2LNOdWFDNA89IB40ca38Q/viewform

[MattS]

Otto

cloud is good but not every customer will use it. There will always be those whose regulatory bodies force them to retain direct control over all of their data b

[Otto]

Thanks Matt. Agree there's lots of regulatory compliance issues to work through. There will be lots of interesting journeys for many customers and a portion of customers will be blocked by a variety of issues. We'll continue to seek feedback from partners and customers in all things cloud so we get the balance right.

[Lars Gronholt]

This makes me nervous if I'm understanding you correctly: that Atlassian's long-term goal is to drop self-hosted options? Should I be looking into alternative solutions then? My employers can't entertain a switch to a cloud-based option.

[Otto]

Hi Lars, we aren't making plans to drop self-hosted. We're interested in hearing from customers as to what their plans are to move to cloud and what might inhibit them. At Summit 2015 many customers seemed to be not only willing and eager, but were asking us to make improvements to our cloud offering to allow them to use it. But like you say, there are also customers that don't want to move to cloud. I'd love to hear from you and others as to the reasons for not moving to cloud.

[Lars Gronholt]

Hi Otto, good to hear! My employers are a government organisation - they have strict requirements on where and how government data is stored due to the sensitive nature of some of it. I'm not aware of the details of the policy concerning data hosted externally, but I believe it has to do with the sensitive nature of much of what we deal with. They simply don't want to risk it being in the hands of anyone but their own people. While the data we (myself and the dev teams using your products here) are moving around is generally innocuous there is always the chance that someone might include sensitive data in ticket detail, a commit, or confluence page. Also, I would expect that if they did - at some point - embrace the future they would probably require assurances that the servers underpinning the cloud solution their data was on must reside in the same country. So that foreign powers would have no capacity to subpoena our data just because a server we may use is based in that other country's jurisdiction (not to mention all the other stuff America's various agencies seem to get away with over there for example).

[Deleted user]

Don't let hearsays let you down :)

[Nic Brough -Adaptavist-]

No.  Whilst I can't answer for Atlassian directly, I have had conversations with their product managers that strongly suggest the complete opposite - they want to support more databases, not fewer. 

In this case, it's pretty much irrelevant what the developers might like to do - there's no way Atlassian would cripple their market by dropping any of the four they currently support.  (Yet.  When the world moves to a better place where weaker databases are not widely used, I can see it happening.   But we're looking at years)

[Otto]

Interesting. I can understand that we would want to support all databases that customers ask for. However, this is a trade off decision. I can clarify that we are *not* looking to add more DBs to our list of supported DBs. Doing so has a fairly significant cost increase (ie. dev hours) that we would rather spend on improving our products. We currently have what would best be described as a "non-proliferation" policy for DBs.

[Nic Brough -Adaptavist-]

I should have said "would like to support more" rather than plain "want". Because customers do want to run it on other stuff (Mongo for example). But compared with the added dev time, I can see why the list is not going to grow.