How to Track Impediments in VSTS

Note: While this post discusses impediments in VSTS, everything mentioned can be applied to TFS as well.

Visual Studio Team Services (or VSTS) has great tools to support Scrum teams. The product owner can use the backlog and board to track the progress of individual teams or the entire product, at the Product Backlog Item (PBI) level, at the Feature level or even the Epic level, throughout the entire lifetime of the product. The developers can track the progress that they are making within the sprint, and see how their work (tasks) fit into the larger picture, by associating them with the PBIs that make up the product.

But what about the Scrum Masters? What tools do they have in VSTS to help them track their work and their progress?

What are Impediments?

According to Scrum.org’s official Scrum guide, one of the services that a Scrum Master provides to the development team, is the removal of impediments to the team’s progress. An impediment is anything that causes a developer to be unable to make progress towards completing the sprint’s goal. Whenever a developer has a problem that cannot be solved within the scope of the team, it is the Scrum Master’s responsibility to remove it.

Impediments in VSTS

Visual Studio Team Services has a work item type dedicated towards tracking impediments, and progress on their removal. For projects using the Scrum Template, this work item type is called an impediment. For projects using the Agile or CMMI templates, this is called an Issue. Regardless of template, they both serve the same purpose: They mention a problem, and their state machine tracks the progress.

Unfortunately, impediments and issues do not show up in VSTS’ backlogs or boards. Those are designed for tracking progress on the delivery of the product, and the Impediment work item type is not included. That said, how should a Scrum Master and the Scrum team track these impediments, especially in large distributed projects, where face-to-face communication and jotting a note on a pad is not a viable solution?

Step 1 – Gathering Impediments

The first step to being able to track impediments in VSTS is obviously to enter the impediments into VSTS. The best way to guarantee that impediments do, in fact, get logged in VSTS, is to make it quick and easy to do so. I suggest using widgets in the Work Dashboard. This should be located prominently in whichever dashboard all the Scrum team members view regularly.

The default dashboard in VSTS (named “Overview”) has a widget titled New Work Item. This widget has a text box for setting the title of the work item, and a drop-down list to select the work item type. You can rearrange the dashboard, if you wish, to make sure that the widget is conveniently located, but otherwise, all you have to do, is select the Impediment (or Issue) type, enter the title (e.g. “I need an MSDN Enterprise license”), and click on the Create button:

clip_image002

This will open a new work item form, where you can add a description, or any other detail that you might want, as you would with any other work item:

clip_image004

Finally, just click Save & Close, or press Ctrl+Enter to save and exit the work item.

Step 2 – Create a Query for Tracking Impediments

As a Scrum Master, I will want to keep track of all the impediments in the team. I will want to note new impediments that are not assigned to anyone for removal, and I will want to keep track of those assigned to me (if there is more than one Scrum Master in the project, which may occur in large projects).

In order to set up your impediments query, you will need to go to the Work | Queries submenu.

clip_image005

At this point you have two options. You can customize the existing Open Impediments (or Open Issues, in the other templates) query, which you will find in the Shared Queries section, under the Current Iteration folder, and customize it, or you can create a new query. If you choose to create a new one, just be sure to save it as a shared query, so that you may use it later with some widgets.

Regardless of whether you’re updating the existing query or creating a new one, make sure you set the following elements:

· Work Item Type should be equal to Impediment (or Issue if in an Agile or CMMI project). This is already set in the existing query

Alternatively, you may make the following variations on the query. You can save these as separate queries, or apply them to the one you are working on:

· You may optionally track only impediments that have not yet been resolved. To do so, you should set State to be equal to Open. This is already set in the Open Impediments query

· You may optionally set a filter to track only the impediments in the current iteration. To do this, set Iteration Path to be equal to @CurrentIteration. This is already set in the existing query, but you will want to modify the value from its current setting which defaults to ‘Your-Project\Sprint 1’

· You may optionally set a filter to track only unassigned impediments. To do so, set “Assigned To” to be empty (the operator should be set to ‘=’ and the value should be empty)

clip_image007

Step 3 – Visualizing Impediments in VSTS

Having a query that shows the list of open impediments is important, but the information must not only exist and be available, it also needs to be accessible. In Scrum parlance, we call this Information Radiation. In this step, we will make sure that this information is in the face of the Scrum Master.

The Dashboard

Depending on how cluttered your Overview (the default) dashboard is, you may want to create a separate dashboard, just for the Scrum Master. Doing so is extremely easy. Just go to the Dashboards section of the project, and click on the + New button on the far right. Give the dashboard a name, e.g. ‘Scrum Master’, and then click on the OK button:

clip_image009

You will now have an empty dashboard that you may fill with widgets. We will use the space to add widgets that help us track impediments. The ‘Add Widget’ sidebar will open (you can also click on the edit button, and then the + button at the bottom right to open it. You may now add widgets.

Adding Widgets

There’s probably no end to the widgets you may add at this point, but I would like to point out the following, as I find them to be of most value:

Query Tile

This widget simply displays the total number of results for a query. Add this widget, and click on the wrench to configure it, as follows:

· Title: Call it ‘Open Impediments’ or something similar. The query name will be the default title

· Default background color: I suggest setting it to blue, or another color that denotes a calm or good state (blue is better than green for accessibility reasons)

· Conditional formatting: Click on the ‘+Add Rule’, and select the red color, with a condition of the number of items being greater than 0

Query Results

This query simply displays the results of a query, as a list. Add this widget, and click on the wrench to configure it, as follows:

· Title: Open Impediments, for example

· Size: You can leave it at the default of 3×2, or whatever size you like, experiment with it

· Query: Choose the ‘Open Impediments’ query

· Display Columns: Choose the columns that you wish to display. I would make sure to have the Title, State, and ‘Assigned To’ columns.

Chart for Work Items

Visualizes the work items with a chart, such as a pie chart etc. You could use this to track impediments over time. This is especially useful with a query that is not filtered by state. You can create a pie chart for impediments, comparing the number of opened and closed impediments, and so on. Add this widget and click on the wrench to configure:

· Title: Open Impediments

· Size: 2×2 is the default, you may change it or leave it, as desired

· Query: Select the open impediments (or a new query for all impediments)

· Chart type: Pie, for example

· Group by: select something to group by, like State, or Iteration Path

· Aggregation: usually Count of Work Items

· Series: Select the color for each group (e.g. red for open, blue for closed)

If you’ve followed my examples, your dashboard may look something like this:

clip_image011

If you can set this up on a monitor that is always on display in a team’s room, this can be a very powerful tool for Scrum Masters.

What other queries and widgets would you suggest for the Scrum Master’s dashboard? Let me know in the comments’ section.

Thanks,
Assaf

Posted in Agile, How To, Productivity, Scrum, VSTS | Tagged , , , , , | Leave a comment

How to Set Up Multi-Level Hierarchies in VSTS

Background

Both VSTS (Team Services) and TFS (Team Foundation Server) have a set organizational hierarchy. An individual developer belongs to a team. A project (also called Team Project, or TP) has multiple teams, and a VSTS account – as well as a TFS collection – contain multiple projects.

While it is possible to run queries in Team Services and TFS that return data from multiple projects, and in TFS, reports that bring data from multiple data sources (i.e. multiple collections), the agile governance tools that VSTS and TFS offer do not aggregate beyond the project level. This means that the hierarchy that VSTS offers has 3 levels: Project, Team, and Developer.

In this post, I will show you how to set up VSTS so that you can create a larger reporting hierarchy, with as many levels as you want.

Teams, Teams, and more Teams!

First, a disclaimer – the following technique, while giving us almost everything that we could need out of multiple hierarchies, cannot create new types of containers or entities. The highest level is still the project, and individual members still belong to teams.

What we can do is set up teams within teams – or at least create the illusion of having done so.

A team in VSTS has the following attributes – it has members (the individual developers), it has its own set of dashboards, its backlogs and boards, and it is assigned an Area under the project.

The way that we specify that a work item is assigned to a certain team, is by specifying that the item’s Area Path is in or under an area assigned to said team.

The trick that we will use to accomplish our goal, is to create teams whose areas are under the area of another team. Each “level” will be a different hierarchical level. We will usually assign the products highest governance (or steering) team to the project root.

For example, if we want our project to have the following hierarchy:

  • Division
  • Group
  • Team

We will create “Division” teams under the project root, “Group” teams under the divisions, and “Team” teams under the groups, as in the following hierarchy:

image

Again, the teams themselves are “flat” there is no team hierarchy. The illusion is created by assigning some teams a default area that is a “parent node” for another team’s area.

In this example, Alpha Group’s area is  MyEnterpriseProduct\Blue Division\Alpha Group, and the Apollo team is MyEnterpriseProduct\Blue Division\Alpha Group\Alpha Team, which is beneath it, but neither team has any other attribute that marks one as higher than another, hence the “illusion”.

But Will it Blend?

So we have successfully created a list of teams, some assigned to areas above others. How do we make sure that the illusion is kept when dealing with boards, backlogs and dashboards?

The trick is to set all but the “leaf” teams (the teams lowest in the hierarchy) to include sub areas, i.e. each team owns its own area and those beneath:

image

Setting the teams like this gives groups a supervisory view of teams, divisions of groups, and the “steering committee” can oversee all of the work being done in the project.

This means that the steering committee’s boards, backlogs, and dashboards will track all of the work being done in the project, while Alpha Group will oversee only the work done by its teams. Each of the “leaf” teams will see the work that has been assigned to them:

Steering Committee's backlog
The Steering Committee’s backlog

image
Alpha Group’s backlog

image
Ares Team’s backlog

This filtering is preserved for the Kanban and Scrum boards as well, and each division, group, and team can have their own set of dashboards to highlight whatever they want to see and use to drive their decision making!

Summary

By creating an Area tree that matches the organizational hierarchy, and assigning teams to their proper nodes, VSTS teams can be made to create a hierarchy as high as the group needs it to be!

I hope you find this useful. If there are any questions, please feel free to ping me in the comments!

– Assaf

Posted in Uncategorized | Leave a comment

Cross-Account Package Management for NuGet in VSTS

In the following post, we will look at the difficulties in consuming packages from a different VSTS account, as part of a build process in our own account.

Background

Visual Studio Team Services (VSTS) offers package management services. As per the documentation, this is “an extension that makes it easy to discover, install and publish packages”. Team Services enables teams to collaborate with other teams through versioned, cohesive libraries and APIs, in the same way that many 3rd party vendors offer their libraries and frameworks today. In fact, VSTS Package Management exposes the most common package management services that developers use today – NuGet for .NET developers, NPM for NodeJS developers, and now in public preview, Java packages can be served as well.

Access to Packages

Unlike the public open feeds, your teams’ package managers are designed to enable access only to those accounts who should have access to the packages. When you create a new feed, VSTS asks you to decide who has permissions to contribute (i.e. publish to the feed), and who has access to read (i.e. download and consume packages from the feed):

image

You have two choices regarding the contribution of packages to the feed. You can specify that any of the team members (i.e. those who are members of the hosting team project) can add packages, and/or the build service account (i.e. the service account used to run builds in VSTS). Note that you will want to specify the latter if you use the build service to publish the packages (you should). You normally wouldn’t need to add permissions for the team members as well.

You also have two choices for controlling access to consumption of the feed as well. You can either limit access so that only members of the project can consume the feed’s packages, or you can allow everyone in the account.

This works fine for most organizations. Not for all, however.

Multi-Account Organizations

Some organizations own multiple VSTS accounts. For some the decision simply grew organically, with various teams trying out the services, and chose to stay this way because there is no easy migration path for a project from one account to another. Others intentionally decided to keep multiple accounts, one for each internal organization (different divisions, for example). Regardless of the reasons, the situation may arise, where components developed in and published by one account need to be consumed by components that are being developed in another account.

Unfortunately, this option does not exist. VSTS does not allow you to specify that you want to give access to members of another account.

How Can We Consume Packages from Another Account’s Feed?

The problem that we are trying to solve is the consumption of packages from Account A, in a build definition that builds an application under Account B. What we will need to do is to overcome the limitations that are set by VSTS’s package management services, and somehow access Account A, with a set of credentials that it will accept.

Solution

Step 1 – Access to the Publishing Account

The first thing you need to do, is to gain access to the account. You need credentials that allow you to view packages. This will be given to you by administrators of that account. You will be given a username, and a Personal Access Token, or PAT (see documentation about creating and using PATs). Your PAT will (ideally) be configured to grant you read access to Packaging for the required account.

Step 2 – Add a Custom NuGet Configuration File

Next, in order to customize how the build system restores NuGet packages, you are going to need to add a custom NuGet configuration file. You can set it up with the defaults, configuring whatever you might need for your own circumstances, like this:

image

Step 3 – Add Your Credentials to the Build Definition

Next, you will need to add the credentials for the package manager’s account to the build definition. You will store these as variables. Be sure to encrypt the PAT – you do not want any passwords to be saved in plaintext!

image

Step 4 – Add a Task to the Build Definition to Access the Package Feed

In the next step, you will need to create a custom source for accessing the other account’s feed. You will need a Command Line build task, and you will be running against the nuget.exe CLI tool. The simplest way to do this is to add nuget.exe to your source control. I prefer to put it under the solution, in a subfolder named Tools.

You will need to call the sources Add command, and specify the name of the feed, its URL, and the username and password/PAT. You will also specify which NuGet configuration file you are going to modify with this setting.

Note that you will want to set the name of the feed to be the same one that you specified in Visual Studio, when building your application.

Here is an example for the Arguments that you must set for the build service to recognize the other account’s feed:

sources Add -name OtherDivision -source https://assaf-account-b.pkgs.visualstudio.com/_packaging/PrimaryFeed/nuget/v3/index.json -user $(NugetUsername) -pass $(NugetPassword) -ConfigFile $(Build.SourcesDirectory)\Custom.nuget.config

Step 5 – Configure NuGet to Restore with Custom Configuration

Finally, you need to modify the NuGet restore command, so that it uses your custom configuration file:

image

Your build definition is now able to restore NuGet packages that are published by another Account!

Summary

In this blog post, you have learned how to configure NuGet in a VSTS build definition so that it can consume packages that are published by a different VSTS account.

I hope you find this useful.

Happy coding,
Assaf.

Posted in Package Management, VSTS | Tagged , , | Leave a comment

Scaling Scrum with Nexus in VSTS

In this post, I will cover what Scrum Nexus is, where, when, and why you would want to use it, and how to set up VSTS best to accommodate your Scrum practices. As VSTS’s tooling is not yet perfect for some of Nexus’s practices, I will discuss some viable fallbacks and workarounds.

Background – Why Scale?

Scrum.org’s official Scrum guide defines Scrum as “A process framework used to manage complex product development”. Scrum’s events, artifacts and rules revolve around a Scrum team, which consists of a Scrum Master, Product owner, and 3-9 development team members (a.k.a. developers).

This limit on team size is important. For Scrum to succeed, the team must consist of developers who can cover all the work required for the product to be delivered at the requisite quality level. If, however, there are fewer than 3 developers, the team is likely to miss some skills required to deliver the product. More than 9 team members, will require too great an effort to coordinate, and result in an unmanageable process.

Therefore, when delivering a large product, as is often the case in enterprise-level projects, the organization must scale beyond the single Scrum team. A framework is required to manage and coordinate the work of multiple Scrum teams.

Enter Nexus.

What is Nexus?

Scrum.org defines Nexus as “an exoskeleton that extends Scrum to guide multiple Scrum team on how they need to work together to deliver working software every Sprint”. While there are other systems out there for scaling Scrum, from a simple (and somewhat naïve) “Scrum of Scrums”, where Scrum Masters get together to coordinate the teams’ interdependencies, to full blown complex frameworks such as SAFe, I tend to prefer working with Nexus, as it is a simple extension of Scrum. It builds upon the knowledge that teams have working with Scrum, and applies the same processes, artifacts, and roles, to a larger scale, introducing minor tweaks, rather than new complex mechanics.

Nexus revolves around the notions that teams minimize risk by minimizing interdependencies. A special integration team (called the Nexus Integration Team, or NIT) is formed. The NIT is responsible to uncover and manage whatever dependencies exist between teams’ work items, by eliminating or carefully controlling their impact. The NIT is also responsible to guide the Scrum teams towards continuously integrating their work, to reduce the risk that comes from large integrations at the end of a sprint or a release.

The Nexus Integration Team members include cross-functional developers, such as the product’s DBA, build-master, architects, technical writer, and anyone who may be of greater use the organization as a coordinator of the integrated product, than a member of one Scrum team. The NIT has one Scrum Master, who depending on the teams, may be the one Scrum Master for the entire product, or just for the NIT. In any case the product must have one, and only one, product owner.

Setting up VSTS for Nexus

The Team Project

First things first – There should be one and only one team project for the entire product. You want to be able to view the entire backlog, measure, query, and track progress for the entire product, and you want to be able to view charts and reports that aggregate data for the entire product. If you separate the product so that each team has its own “Team Project”, you will not be able to do so. The Work tracking and dashboard capabilities of VSTS are limited by scope to a single Team Project. You cannot split a backlog, or visualize a Kanban board across multiple team projects.

This choice does have some limitations – the entire product, including all of its teams will have to follow the same process template, though if you have already decided to follow Scrum and Nexus, this should not be an issue.

Areas and Iterations

One Set of Iterations for All Teams

According to the Scrum guide, if you have multiple teams working on the same product, they should be delivering together on the same cadence. In VSTS, this means that there should be a single iteration cadence that all teams follow. All teams start together and end together because the integrated product increment is delivered at the end of every sprint.

The Nexus Integration Team is the Default Team

In VSTS the default project team has the project root as its area. Each of the other teams get an area under the project root:

clip_image001

The NIT Includes all Sub-Areas

In Nexus, there is a unique event that is added to all other events that normally occur in Scrum. This event is called the Nexus Sprint Planning session. In this meeting, the product owner and the NIT review the upcoming work and coordinate what PBI will be addressed in which sprint, by which team. In addition, the Nexus Integration Team is responsible for coordinating of the integration of all teams’ work, on a daily basis.

In order to accommodate these needs, the NIT’s backlogs and boards will be configured to view not only the work items (Epics, features, and/or PBIs) that are in the NIT’s area (the root area for the project), but also the work in each of the areas beneath it, i.e. work that has been associated with the Scrum teams.

This way, the integration team can visualize all of the PBIs on their board, and be ready whenever a PBI is ready for integration (e.g. moved to a column entitled “Integrate” – you will want to create this column).

Setting Up the Nexus Sprint Backlog

As previously mentioned, one of the NIT’s primary responsibilities is to coordinate work among the Scrum teams in order to mitigate interdependency risks. A commonly used tool is called the Nexus Sprint Backlog. The Nexus backlog is views the PBIs in a two-dimensional grid, where they look at each team’s sprint backlog for the upcoming and next few sprints. They then identify and mark each dependency, and note how risky it is:

· When interdependent PBIs are handled by different teams (e.g. 1 & 5, below), the risk is greater than when both PBIs are handled by the same team (e.g. 1 & 4) because coordination is more complex.

· When PBIs are not only handled by different teams, they also are expected to be completed in the same sprint (e.g. 4 & 5), the urgency requires greater coordination, and the risk is even higher.

· When there is a dependency on a PBI by someone outside the project, the risk is greater because there is likely to be an important commitment (e.g. 8). This risk goes up even further, if the PBI is needed by the end of this sprint (e.g. 9)!

image

While there is no Nexus Sprint Backlog in VSTS, as of the writing of this post, Microsoft’s Delivery Plans extension goes a long way towards giving us just this visualization. It is available in the marketplace so you can install it and use it to visualize your work across teams and sprints:

image

Visualizing Dependencies in VSTS

The delivery plans extension does not have the ability to visualize and track dependencies between backlog items. Yet. The product group has mentioned that this capability is on the roadmap, but I cannot confirm when this will be delivered.

In the meantime, we need to come up with a way to mitigate this drawback.

Work Item Relationship Links

One thing that you can do is mark a work item to be the predecessor of another. To do this, simply open the PBI, and under Related Work, click Add link to an existing work item. Select the work item that this work item depends upon:

image

While this will set the work items’ interdependencies, this will not show up in the plan. At least not as of today.

Tags

While this requires some manual work, I would suggest adding tags marking an item as “cross team dependent” or “external dependency”. Tags do show up in the plan and can augment it to give the NIT an at-a-glance idea of the risk involved with this project. Note the tags marked by the arrows in the screenshot below:

image

Summary

We have seen how teams wishing to use Scrum and Nexus to drive their development efforts, can set up VSTS to visualize and track their work. The following is a checklist for everything that you need to do:

1. All teams including the integration team work in the same Team Project

2. The Integration Team is the default team

3. The integration team’s area is the project root (e.g. \MyProject)

4. The Scrum teams’ areas are set under the integration team’s (e.g. under \MyProject\ScrumTeam1)

5. The integration team’s area should include all sub-areas

6. Add a column board entitled “Integrate” to mark PBIs that are ready to be integrated with the entire product

7. Install the Delivery Plan extension if it isn’t already installed for your account

8. Set up a delivery plan to include the Scrum teams and their sprints

9. Use the Predecessor link type to denote a dependency

10. Mark the kind of dependency with a tag, so that it can be seen in the plan.

I hope you find this useful. If you have found other tip and ideas to help drive Nexus, please share them with us!

Thanks,
Assaf

Posted in Agile, Scrum, Team, User Story, VSTS | Tagged , , , , , , | 1 Comment