Jira Day 2018 was the biggest ever Atlassian-related conference in Central Europe. 380 guests, 18 speakers, 13 Partners and vendors from all over Europe gathered at the Legia Stadium in Warsaw to share their expertise and discuss the latest trends in development of the software. We’ve recently started passing their wisdom on to you, and in this article we’ll cover the first part of the prominent topics touched upon at the event.
Agile Project Management
After the official event opening, when our CEO Piotr Dorosz proclaimed the company’s recent rebranding, the keynote was spoken by Marcin Żmigrodzki, who is a project management expert helping companies and teams work effectively by applying innovative techniques such as simulation games. Marcin presented the recent Google research results that revealed 5 key factors to building an effective team. These are: psychological safety, dependability, work structure and clarity, along with the feeling of personal meaning or impact on the company’s business results.
‘Who is on a team matters less than how the team members interact, structure their work, and view their contributions’, the speaker concluded.
During the interview that Marcin gave us after his speech, he extended this idea by stressing that every team member must be informed about what’s going on the project to be effective. Among those “informing” techniques, retrospectives and task boards appeared, which are pretty familiar to those working by an Agile or Scrum-inspired model. Asked about the advantages of Agile project management, the expert listed independence and decisiveness of teams, as well as increased motivation compared to the Waterfall model. However, he mentioned reduced control possibilities, dependence on individual skills and the necessity to make collective decisions as the drawbacks of Agile. ‘It’s always good to consciously choose management methodologies, not just follow trends’, Marcin admitted.
After the keynote, Atlassian Senior Product Manager Jakub Kurcek went on stage. He presented their recent changes of Jira Software, difference between different deployment options (such as Server, Cloud and Data Center), and their development plans for the rest of this year. The main improvement areas are the user experience and self-administration capabilities of the software.
‘Jira is heading to be more Agile than ever and improve UX dramatically’, the speaker claimed.
More configuration capabilities for the project managers have been added to Jira – now they can create fields, issue types and workflows or invite users without engaging a Jira administrator. The new Agility board has been introduced to Jira Cloud, which we’ve already mentioned in App Week 2018 summary. Basically, any user of the board can now quickly create and move columns on the fly. This means that you can edit the basic workflow directly from the board, so there’s no need for a dedicated editor to apply basic changes. Teams that are starting their Agile journey can easily adjust the board to their changing workflow, starting simple and increasing the complexity along the way. If these functionalities seem a bit familiar to you, remember that last year Atlassian bought Trello, so these are pretty obviously inspired by this new acquisition. The issue detail screen was redesigned completely as well to improve its usability. Now it shows only the information users need, selectively hiding redundant fields. ‘This is the area where your team members spend most of the time, so the simplified view means way more signal and way less noise’, Jakub said. Also, the brand new quick search, which was developed after Atlassian had moved their whole infrastructure to Amazon Web Services, now searches projects, boards, issues, and filters from anywhere within your Jira.
These new features, as well as the upcoming ones, are deployed on Cloud first. The reason is that Cloud instances are upgraded automatically by Atlassian, so they can test it faster and shorten the feedback cycle. Due to hosting specifics, the roadmaps for Server, Cloud and Data Center are somewhat different, but the key improvements will appear on all of them over time, as the Atlassian ensured.
Atlassian Experts Bogdan Górka and Przemysław Gochna presented their concept of so-called “fast moving projects“. They compared Agile development with the FMCG branch, as both cases are characterized by short delivery time, predictable project flow and frequent releasing at a relatively low cost. The speakers listed the common problems of such projects, like email abusing, constant change and information silos. To solve and prevent them, they recommended Kanban boards and thought-over workflows in Jira, along with Confluence as a unified knowledge base.
Michał Wachnik, Program Manager of the Polish telecommunications company Polkomtel, revealed an interesting case of how they customized Jira to create a fully functional, flexible and ergonomic project management tool, used now by 750 people in their organisation. It took them 5 months to accomplish, and they remain satisfied with low costs of both implementation and licensing, seamless integration with the development tools and possibilities of further customization, as well as scaling up the solution.
App vendors brought their own solutions for project management, too. Guillermo Montoya and David Garcia from DEISER endorsed the Profields app, which offers more fields for projects (like budgets and due dates), the Project Navigator for handling multiple projects, along with enhanced reporting capabilities in Jira. Eyglo Oskarsdottir from Tempo presented an opportunity for Project Portfolio Management in Jira with a single, yet expanded tool. Tempo Timesheets, Tempo Budget and Tempo Planner cover all the needed areas of PPM, such as Pipeline Management, Financial Management, Resource Management or Change Management. In his turn, Martins Vanags brought the well-known eazyBI tool from Latvia, which is a powerful business intelligence tool for data analysis and visualisation inside Jira.
Scaling Atlassian Solutions
Recently, more and more companies have faced the necessity to scale up their Agile practices. More teams at big enterprises are willing to adopt the methodology over time, and smaller companies often find themselves growing rapidly. Therefore a variety of scaled Agile frameworks have emerged, and having thousands of users and millions of issues in Jira is not so uncommon anymore. The latest report from Atlassian states that an average enterprise customer’s instance counts 1,4 million issues per year, 1200 concurrent projects and 400 workflows, which is quite impressive. This high amount of data and complexity level can cause performance issues and various problems for Jira administrators, as they should configure the solution properly and ensure further scaling opportunities. Stable data flow is also becoming very important, as more teams are geographically distributed around the world or include remote workers. So this was definitely one of the hottest topics on the event’s agenda.
Jakub Kurcek stated during his speech that scaling organisations requires improving tools to administer and manage all teams’ efforts. What’s more, the tools have to be scalable along with the organisations that use them to operate and manage work effectively. That’s why Atlassian products are now offered in three deployment options: Server, Cloud and Data Center. The latter is the newest offering by Atlassian, which is designed specifically for enterprises operating at scale. With this option, you can run Atlassian software in a cluster inside your own environment, on your own hardware or a private cloud, on a database and operating system of your choice. As Jakub pointed out in the interview, Data Center hosting features increased performance in case of high data flow and big number of users, better availability with less downtime and tighter user database control. After all Atlassian infrasctructure had moved to AWS, Cloud became another good option for scaling up Jira, as there’s no need to add up hardware nor engage system administrators to upgrade an instance. However, the Server hosting will still be supported and enriched with new features.
A practical case of Jira Data Center implementation at an enterprise was presented by Sebastian Krzewiński, Senior Application Engineer at Allegro. The company is the biggest e-commerce marketplace in Poland, having 70 million listed items, over 20 million registered users and 14 million unique visits each month. At a certain moment their Jira Server implementation could not handle 2000 users across 5 locations and 950 projects with 2 million ticket volume anymore, so they migrated to Data Center to address the situation. The overall goal was to eliminate downtime while upgrading or re-indexing, as well as retain high performance of the setup. Sebastian went through the migration process and provided solutions to the technical problems encountered. The result achieved was basically a two-node architecture, which made Jira available to users at any moment by checking endpoint /status every time a user logged in and then using a persistence profile to keep this user on the running node.
Then Atlassian Architect Tarun Sapra presented the way that scaled Agile operations are executed at Ingenico, the company he is currently working for. They’ve adopted a variation of the Spotify model, described by Henrik Kniberg and Anders Ivarsson in a white paper, which allows to keep an agile mindset despite having a multitude of teams in various locations. During his presentation, Tarun revealed the key success factors of this approach, which were proper requirements management, smart workflow setup and process automation.
‘Working with 200 teams in a scaled Agile environment, we need to map the process with the tools. Requirements have to be collected in a centralized location, given structure and achieved traceability’, the expert stated.
It turned out that Atlassian solutions, set up and customized with apps properly, are an ultimate answer to this challenge. The speaker particularly stressed the importance of reporting and linking requirements with the features, having called out Confluence, ScriptRunner and eazyBI among the most useful tools to help accomplish it.
Another example of scaled up Jira was explained by Mirosław Bartecki, Chief Architect at Capgemini Financial Services. Their customer support solution operates globally 24/7, with 7 million tickets per year, 1 million attachments, 8000 concurrent users on average, 1 TB full text index and 110 integration endpoints. It’s easy to guess that such high data amount would inevitably cause performance issues with Jira. As the speaker explained, the biggest challenges came from the small components of the software, such as customized workflow steps, text indexing and attachments. At a certain point, one click in Jira could produce up to 2 million API calls, which along with the global distribution caused significant delays in the data flow. The best practices for performance tuning included: deep diving in analyzing KPIs, a test instance with 100% data volume and test case covered, and plugin optimization to reduce the risk of errors. ‘It’s worthwhile to collaborate with Atlassian Solution Partners and app vendors on this matter’, Mirosław admitted.
One of the apps helping leverage Agile operations at scale is Structure for Jira by ALM Works. It enables keeping track of issues from any project, folder or Confluence page, also featuring a custom API to integrate with other tools (Xray, Portfolio and Tempo, to name a few). Use cases include scaled Agile frameworks, workload balancing, personal workplace or test case management. ALM Works SAFe Expert Julia Atlygina demonstrated the app’s capabilities and provided pro tips to set up such an environment, which included creating Jira projects per level and not per team, use of Scrum boards, workflow setup and links between items in the Atlassian Stack.
Atlassian Stack Integration
The Stack was introduced by Atlassian as an all-around work management solution in June 2017. At the time, almost half of the customers with 500 or more users owned 3 or more Atlassian products simultaneously due to their close integration possibilities. So the makers of the software decided to integrate those products even closer with a goal of fully supporting DevOps in companies with their set of tools.
Our own Atlassian Expert Piotr Mazij presented the details of what the Atlassian Stack usually consists of and the wide array of activities it can be used for. These activities cover project planning, task processing, customer support, content management, team collaboration and all kinds of work with the code, from peer reviewing to final deployment. The main tools of the Stack are Jira Core, Jira Software, Jira Service Desk, Confluence and Bitbucket, but there are many more in the Atlassian product line. They stand along with over 5000 Marketplace apps to help companies create flexible and fully integrated environments for managing all their work.
Tarun Sapra mentioned the integration of Jira and Confluence as a crucial part of successful requirements management using Atlassian Stack. ‘If everything is done right, product owners even don’t have to open Jira, as they see everything in Confluence itself’, the expert said. In order to do this, he also suggested exploring REST API and apps like ScriptRunner to automate creating charts or release notes from Jira and publishing them to Confluence.
Jira Service Desk Customization
At first sight, Jira Service Desk may seem to many as rigid and lacking features out of the box. But it is actually highly customizable — the secret is in the apps from the Atlassian Marketplace, which help adjust the tool to companies’ needs and provide excellent customer support. The real-life cases presented at Jira Day 2018 prove efficiency and cost-effectiveness of such combinations for IT service management.
Joanna Czochara, MetLife’s Computer Support Supervisor, shared their history of pioneering Jira Service Desk implementation in Poland. Before migrating to the Atlassian solution, they had used several systems and processes to manage requests, including email, open source OTRS, IBM ClearQuest or even paperwork. This situation was uncomfortable both for the IT department and for the users, so the decision was obvious when they noticed Jira Service Desk hitting the market in October 2013. Immediately after establishing a test environment, they felt the need to customize the solution and asked us to develop some features for them. Those were Dynamic Forms, request type security and a bunch of others, which basically started the history of Extension for Jira Service Desk. Another area of improvement was integration with Active Directory — Catalogue Synchro and Application Synchro, which are 100% custom features, helped MetLife resolve 25% of access management tickets automatically.
Then our own Senior Atlassian Apps Manager Kate Pawlak pointed out the most common concerns of Jira Service Desk users — both agents and customers. They are mostly related to the request view on the Customer Portal, filling in long forms and managing queues.
‘We’ve been observing the use of the Atlassian tool for ITSM over the years and gathered enough feedback from companies to develop 4 apps that address these pain points: Extension, Actions, Queues and Translation’, Kate shared.
With their help, customers can track SLAs for their requests, see attachments and issue links on the request view and have the Customer Portal translated into their mother tongue. On the other side, service operations managers and support teams can have enhanced cross-project queues, the ability to create dynamic forms and add extra information to the issues on transitions.
Another app that helps customize Jira Service Desk, as well as Confluence, is RefinedTheme by RefinedWiki, one of our Partners. This one deals with UX and appearance of the solution, allowing to create clear, engaging and easy to navigate Customer Portals, tailor them to user groups and apply a company’s brand to their appearance. Talking about the theory behind the app, Genevieve Blanch mentioned the key factors to take into account when building a perfect homepage.
‘The context should be perfectly clear to anyone who sees the page, and they should be able to find answers to their questions quickly. The experience should be engaging, too — this is actually a great indicator that the page is built correctly’, Genevieve explained.
Data Security in Atlassian Solutions
For companies dealing with sensitive customer data, data security is crucial to handle service operations properly. It requires thought-over permission schemes and workflow setup to ensure that there won’t be any leakage, especially in big enterprises with thousands of users and millions of issues in their Jira. Experts gathered at Jira Day 2018 provided pro tips and tricks to maintain security in Atlassian software.
Katarzyna Olchawa, Senior IT Engineer at mBank, showed a multitude of ways to set up approvals in Jira, including workflow conditions, sub-tasks, custom issue types or dedicated apps. Some of these methods require additional tools like Jira Workflow Toolbox, ScriptRunner or our own Approvals app, which places acceptance transitions onto issue statuses automatically and allows to customize them. The possibilities for setting up acceptance process in Jira are really wide, as the actual choice depends on what’s needed. Accepters may be chosen manually or automatically among project roles, user groups or individual users, the votes may be submitted in order or simultaneously, etc.
Katarzyna also provided guests with additional hints to better manage approvable tasks, such as creating queues of issues waiting for acceptance, building Agile boards including the approval stages, or having dashboard gadgets set up for the accepters. Another good practice is setting up additional transition screens for rejecting, where the voter can provide reasons for such a decision.
Talking about sensitive customer data in Jira during the presentation, Mirosław Bartecki admitted that financial data security was a serious headache for Capgemini, as they needed to have a custom workflow for each project role per client and retain Jira Service Desk performance at the same time.
By implementing the so-called ‘fine grained security’ approach, the expert managed to downgrade initial 40,000 roles they had on their Coca-Cola Customer Portal to less than 100.
The main assumption was that one cannot require acceptance on each workflow step, so the person who started an activity will see all the data related to it. This approach allowed to focus on restricting cross-tenant data visibility and deploy a balanced security model.
Jira is not the only Atlassian tool which requires thought-over permissions, though. Gergely Gurmai and Laszlo Sziacs from META-INF showed that clearly with a puzzle game for Confluence permissions, which was actually fun to complete. Their Ultimate Permission Manager solves the problem of extensive permission schemes for spaces and pages in a knowledge base. Users can easily check who can view a particular page with a dedicated macro, space admins have got the Effective User Permissions screen which allows to build permission schemes faster, and Confluence admins can use the Central Permission Overview to effectively check and manage permissions across all the spaces.
Jira Automation and Machine Learning Opportunities
As more big corporations adopt Agile transformation over time, the problem of workflow complexity arises more urgently for them. Teams having extended workflows naturally start searching for ways to automate repetitive actions which take much time to complete manually. These can be simple workflow actions inside Jira or tasks related to its integration with the rest of the Stack and external tools.
Fortunately, Jira provides a couple of automation capabilities, such as REST API calls and groovy scripts implementation. They often require coding skills to build custom features, but an experienced Jira admin can make miracles with those methods.
Quite often, Marketplace apps provide automation for default Jira features as well. A great example came from Maciej Gajownik, founder of CoqIT, which is an IT infrastructure and services company. Trying to deal with ordering and controlling virtual machines for his clients, he found an extremely powerful combo of Jira Service Desk with Insight, an app made by Riada. Having the entire CMDB stored in Insight, Maciej could create structures of data and metadata with one-to-many references between objects: tech specs and logical roles of the gear, contacts to customers or technical people. Then, using Insight’s dedicated custom fields and post functions, he automated both creating Insight objects based on data put in during request submission and autocompleting Insight fields based on object choice or other Jira Service Desk fields’ values. The next step was to automate configurations’ deployment for these servers. The Insight app supports this automation for Bamboo, Puppet, Chef, AWS CodeDeploy and JetBrains TeamCity. The setup included mapping configurations between Insight and the deployment system 1:1 and write REST API hooks to grab them from one another.
An example of automated integration within the Atlassian Stack was shown by Atlassian Expert Maciej Ląd, who had implemented automatic RFC creation upon running a web service for his client. This required integrating Bamboo into the Stack and implementing groovy scripts deployed back to Jira.
Another automation tool for the Atlassian Stack is Issue Publisher by catworkx. It is basically a post function that automates creation of Confluence pages from Jira issues, which is very useful for generating project documentation, contracts, offers, invoices, “What’s new” pages, etc. Additional features include copying attachments and overwriting Jira field values in a dedicated macro.
What’s even more interesting, Mirosław Bartecki is currently exploring machine learning possibilities for Jira. He told us about the innovative Service Desk automation that he is working on for Capgemini. Their support system operates globally for 50 big clients, who are distributed geographically, so the problem of redirecting tickets between different locations and teams arose quickly. Mirosław develops a solution which redirects tickets to the appropriate teams based on the context of the ticket: location, language or semantics of the sentence. Even though similar effect can be achieved with Jira Service Desk rules available by default, the keywords should be explicitly present in the request, which is not always the case.
‘A machine learning system embedded into Jira is able to recognise a user’s problem and location precisely from the context, so that less errors occur, and the whole process runs automatically’, the expert concluded.
Such innovation is warmly welcome by Atlassian. Jakub Kurcek even said in his interview that Jira may become an artificial intelligence itself in the future: ‘It may happen so that instead of helping us doing things, Jira will actually do things for us’. Time will tell, but it sounds optimistic — hopefully we’ll hear more on this topic at Jira Day 2019!