Unleashed 2023

5 minute read
What is DevOps?

What is DevOps?

DevOps is the continuous delivery of prioritised value to an organisation. For more click here.

Last week (December 14th 2023) I was in Amsterdam for a couple of days with Atlassian product managers and here follows some impressions;

Several products that were in beta have now achieved GA and of these Jira Product Discovery appears to be the star turn. Compass and Atlas are clearer in their positioning but there is acknowledged overlap between all of these products. The emerging challenge for our customers is matching specific users’ problems to the correct combination of products, product editions and add-ons and the scope of that task goes beyond just Atlassian. Thanks to ‘try and buy’ you can at least play with pretty much everything described below.

The over arching theme for the event was AI.

Atlassian has now embedded AI into its core platform (i.e.across all products) and in a world where according to Elon Musk data is the new gold – Atlassian has circa 20 years of project issues to inform its Small Language Models (SLM) which are ‘backed’ into OpenAI.

One of the more interesting audience interactions was between a product manager who asked the attendees how many were exploring AI and 90% of the hands were raised but when asked if there was a specific use case – only one or two people could raise their hand. This summed up much of what now follows – AI within the Atlassian tools are solutions in search of problems. Hence the need for Jira admins to get to grips with exactly what the dev, ops and DevOps and business teams really need – not what they say they want – but what is the problem statement and can you feel their pain?

That said AI Is now free of charge for Premium and Enterprise customers so there is good reason to explore what is on offer. For those on Standard – upgrade to premium for a couple of weeks and take a peak.

Some of the worthier (IMV) features are ‘old school’ AI – so natural language inputs to output JQL and Automations. Also given LLM are best suited to text – it’s no surprise Confluence has AI powered new features – and for those who are thinking – ‘so what – my Confluence pages/spaces are a mess’ – there is hope.

Confluence new features include the AI editor, a drop down ‘prompt’ list (the list is editable) and your Agent will take the text you have inputted and clean it up. For example summarise your text or rewrite it for better readability. You can in turn use this AI tool to clean up the garbage in your Space and automate tasks such as archiving pages that have not been viewed in 6 months etc. Or find and change common acronyms or program names or update dates across all spaces. It may not seem like much – but these should add up to decent individual productivity gains for those that spend significant time in Confluence.

In both JS and JSM there will be Agents embedded into workflows to suggest how best to undertake tasks. It was openly stated by the most senior product manager that there is a deliberate effort to be more opinionated and proscriptive about ‘what good looks like’ for process but that is not yet something with a committed delivery date. An example of where this could apply at some future date would be to recommend a workflow with fewer transitions or suggest an automation. Within JSM today virtual Agents can improve deflection or summarize requests for human Agents.

One upside to the large volumes of data in the Atlassian cloud is that you can expect ML insights and predictive decisioning and this was evident in the Atlassian Analytics product, pieces of which also surfaces in Compass. (see below).

Atlassian are at pains to point out their AI products respects your user permission – prompts will return different data to the same query based on your role in the organisation. Also your data is not used to train other clients. AI is an opt in per product and this is provided by the org admin. All of this is available at no cost in the premium and enterprise editions and the recommendation is to turn it on in the sandbox in the first instance. Finally it useful to know that not all updates to Open AI automatically roll into the Atlassian SLM.

Jira Product Discovery (JPD)

As stated above JPD is the star turn. It has moved from Beta into production 6 months ago and has already attracted 4,000 new customers and an even larger number are currently trialing. Atlassian assert that they have tapped into a change from project to product that now helps a Project Manager to keep a handle on the details but with an always on connection to the ‘why’ of a product.

With the ability to bring in bug data from JSM and run surveys in Confluence or ideation workshops in one central location the hard data is presented to help Product Managers to abandon ‘gut’ instinct. The tool is particularly useful for managing ‘maybe’ or ‘uncommitted’ work, helping teams what to prioritise before deciding what should be moved into Jira and through to production. Here is what one local customer told me…

I think JPD is an excellent product, but what makes it great is that it’s Jira without all the heavy jira baggage. No hierarchy, nice visuals, and some really nice tools to gather feedback, evaluate priorities in a priority framework, and make priority decisions based on data. Our PMs like it because it gives them a sort of place to play without triggering engineering by creating Jira tickets

Atlas

Atlas is essentially a team coordination and information aggregation tool. It’s designed to centralise the information about teams, their ongoing projects, and the tools they use. Teams can be business or technical and this product is particularly useful for geographically dispersed or remote work. In fact the tool was conceived during COVID.

Functionally, it offers a platform for tracking team goals and project progress, acting as an informational hub where updates are shared.

Key features include goal tracking, providing brief updates (akin to ‘tweet-sized’ updates), and linking these updates to specific team goals or projects so that as you work with folks not sitting beside you – their work and yours can stay aligned.

Compass

This tool is aimed at larger teams – probably teams with interdependencies between core platform software and functional components where the need to be aware of changes to interdependencies traverses different time zones.

Compass centralizes your software architecture, providing a single point of access to all your components. It essentially serves as mission control for your distributed system, streamlining the process of managing microservices, libraries, and other components in one unified place. It also offers tools to track software health metrics and apply security and health scorecards. e.g.DORA metrics.

It will improve your component management to give full visibility into service details including APIs, libraries, documentation, key health metrics, recent deployment activities, on-call schedules, upstream and downstream dependencies.

The platform enables tracking of DevEx (e.g. DORA) metrics across teams and services to identify bottlenecks and improve the developer experience

Finally Compass facilitates the creation of templates, which automate infrastructure provisioning and deploy pipeline setup, incorporating your own bespoke observability, service details, and security automatically.

Appfire

If you have a number of add-ons chances are Appfire is one of your providers but you may not be aware of this fact. Quite a few of our customers have JSU and JMWE – and these are now part of the Appfire stable. Appfire have been consolidating Atlassian add-ons over the last several years and have grouped them together into 4 solution packs;

IT Service Management
Manage and Migrate
Workflow and Automation
Agile Work Management

And the main topic of their presentation was Agile Work Management. This suite of add-ons includes;

Agile Poker for planning and estimation
Foxly a prioritisation aide
Big Picture for program management
Dashboard Hub for Jira to visualise work for Agile and DevOps teams
OKR for Jira – if Objective Key Results are your thing
Hedge – a risk management tool for Jira, designed to identify, categorize, and mitigate risks.

Appfire for this suite of tools is aimed at organisations who struggle to see the value of the Premium edition Jira product. They have pulled these products together as solution packages – but do not at present sell them as a bundle.

In other news….

Just when you thought Atlassian could no longer confuse you with a product name change – guess what?. Advanced Roadmaps is now renamed Plans. For those of you old enough to remember when it was called Jira Portfolio Management take a bow.

Jira Align is no longer a slave to SAFE and indeed even those SAFE deniers amongst you will be relieved to hear that constraints imposed by a SAFE implementation have now been removed.


POSTED IN: DevOps Best Practice, DevOps tools, Pipeline automation, Uncategorized