We're a small software team that has been using Jira Server for many years.
During the last two years we have explored Cloud migration a few times. This year we started migration with the hope of completing this in a couple of months. Server support ends in 2024
We encountered many problems, and after 6 months are closing the few remaining migration blockers. It seems to me :
For others that have used Jira Server; I'm curious to know:
Hi Rilwan - thanks for the comment.
Yes, changing from Server to DC would be simple. However the minimum Jira DC license tier is 500 users (US$42,000/year) this is not an option for small teams.
Have you already migrated from Server to Cloud, or are you using Jira DC?
For others that have used Jira Server; I'm curious to know:
- Have you already migrated to Cloud? How long did it take? Are you happy?
- Are you planning to continue using Server (or Data Center)? Why?
- Is anyone planning to replace Jira with something else?
1/ Yes i have migrated a lot of customers to Cloud or to Data Center. They are all happy in Cloud or in Data Center ;) For planning and time: It depend mostly on the number of used apps, the complexity of all configs, and the number of projects, issues and attachments, it can goes from 3 days to 40 days (and more) for a migration to be done. Data Center versions are made for medium or big organization that wants to scale the usage of Atlassian products building a cluster of servers to have better performance and also new features.
2/ No all my customer's Jira are exposed on internet so staying on Server with no maintenance is not an option for security reasons.
3/ No all my customers stayed on Jira cause Jira is really well integrated with others Atlassian products such as Confluence, Bitbucket, Bamboo, and others... Few customers studied others tools but because of the integration capacities, the features in Jira and the cost of migrating to another tool, they all stayed on the Atlassian stack.
I hope this helps.
Cheers,
Dam.
We moved from Server to Datacenter a couple years back. We're larger than what you describe, so having multiple servers including a dedicated API server has been good for us.
We have (and will continue) to evaluate Cloud. It used to be that Cloud just wasn't feasible for feature and scalability reasons. That's past now (mostly), but Cloud also doesn't have any great features that we feel we're missing by being on Datacenter.
However, since Cloud is an all new API for third parties, we see that major add-ons that we rely on (Adaptavist ScriptRunner, ALMWorks Structure) are not at feature parity in Cloud, and our own in-house add-ons would also need to be rewritten.
Until that balance changes, we don't see a strong motivation to move to Cloud.
Your comment about performance in NZ is interesting to me too, since our teams are spread globally. I would have thought that as an Australian homed company, Atlassian's performance in it's own part of the globe would be stellar.
I think the point about performance in NZ is probably under-explained here. We don't know where the main set of users are for this system, and the performance might be down to simple latency. If, for example, the Cloud system was owned by a French bank, which has a small office in NZ, then it's likely that the Cloud system is running in Frankfurt, a long long way from the NZ users.
@Nic Brough -Adaptavist- is there a multi-home capability? If we moved to Cloud and had to choose a single location then that'd be good for our US East Coast people, but sounds like it'd be terrible for folks on the other side of the globe.
And while you're here - anything you can say about when ScriptRunner will reach feature parity on Cloud?
No, there's no multi-home. The only way to do that would be to have two separate Cloud instances.
SR for Cloud is never going to have feature parity with Server. Cloud is a very different platform, and there are some things that cannot be done on Cloud the way they are on server. SR for Cloud is in fact, a complete rewrite - there's no code from the server version in it, just some of the ideas.
(But we have managed to get Atlassian to add extension points to Cloud so that we can do something like Behaviours at last - that functionality is in Beta right now)
Thanks @Nic Brough -Adaptavist- - this is key to our Cloud migration considerations. What's the best way to check on specific SR functionality and whether it is in the "ain't gonna happen" bucket or the "we're working on it" bucket?
Well, there's the docs at https://docs.adaptavist.com/sr4jc/latest - actually if you look through the "migration" section, you'll see a lot of what you can and can not migrate.
But a better answer is probably to fire up a Cloud system, throw Scriptrunner on it and test for the things you currently do with it on Server.
Your comment about performance in NZ is interesting to me too, since our teams are spread globally. I would have thought that as an Australian homed company, Atlassian's performance in it's own part of the globe would be stellar.
Good question @John Dunkelberg - to elaborate my note about performance from NZ...
After using Jira Server for many years, we found Jira Cloud to be very sluggish when we first explored migrating (a couple of years ago). It is clear that Atlassian were well aware of this and have worked to address performance over many years.
Our current Jira migration testing project started in 2022. Again we found Jira Cloud sluggish in comparison to Jira Server (though better than before). After raising this with Atlassian I am unsure why this is the case. Our users work in the same geographic location (near Wellington). We use modern PCs, on a fast local network, and have a high-speed fiber broadband connection (950Mps).
I wonder whether our internet connectivity to the nearest Atlassian Cloud DC (Sydney) is a factor in this. Yes we're in the same part of the world as Atlassian in Australia, but:
The end result is that the latency and bandwidth we get to the Atlassian DC must have some impact on the performance an end-user experience. This seems to vary, and apart from making the product more efficient, it is not something Atlassian can fix.
We hope that the slated performance improvements for Q2 2022 help. The last planned performance improvements are due Q4 2023, just before Jira Server support ends.
Thanks @Dwight Holman - that's good information and a clear hurdle that Atlassian needs to remove before people can move off of on-premise solutions.