Continues Integration and Continuous Delivery with Azure DevOps - Deploy Anything to Anywhere with Azure DevOps

Editor's Notes

• #3: Few words about me
• #4: So set start with, what is Continuous integration First time use of meaning “Continuous integration” in 1991 year
• #5: So let talk what is CD now.
• #6: What is DevOps. Quota Ci, CD and DevOps are path of Agile approach.
• #7: So what is Waste And this is connected to ->
• #8: Lean software development / management / manufacturing And we have translation of seven waste from Manufacturing to Software Development Partially done work Have you ever written a feature that didn’t have fully fleshed out requirements because of a lack of grooming? Have you ever written code and found that the feature clearly needs more work but by definition you’ve completed the feature? If so, these are a perfect examples of partially done work. They require us to go back and ask the Product Owner to go back to the stakeholders and rethink what they asked for. It introduces testing challenges such as incomplete unit tests. It’s also difficult to measure the impact of the feature if we have to repeatedly go back and ask for more details. Extra Processes Do you have to get management approval before creating a new feature? Do you have to get a sign-off before a new release? Does your cyber team have to review the code before it is released? Do you have to document unnecessarily? Some of these questions may seem like they are absolutely necessary. For example, we need to have our code reviewed for security vulnerabilities, right? Of course, but do we need to have another team do this? Could we not have the cyber team teach developers how to write secure code so that they don’t have to wait? Couldn’t we automate this process with some tool as part of our build process? Oh yeah! Extra Features Ever agreed to work on something from the backlog and had the creative juices start flowing and find that before you know it you’ve added extra features that weren’t originally called for? I certainly have. It’s part of the creative act of software development. Even though it is fun it can also lead to waste because it isn’t what the customer asked for. It leads to extra testing, extra builds, extra releases and potentially extra bugs and defects. Gold plating is a form of waste and no matter how tempting it is we should avoid it. Task Switching Ever been working on something to be only be interrupted with another request? Yep, happens all the time in most organizations? Often IT professionals are asked to work on not just two projects but in some cases as many as a dozen. This is a waste. Every time a manager or project manager comes asks someone to stop what they’re doing and turn their attention to a higher priority item it requires a context switch. That context switch makes the person stop what they were working on and begin working on something else. We can control this in agile planning by setting our Work in Progress. See here. We can also self-organize our teams as I’ve shown here. Waiting Have you ever waited on another team to complete their part of something you needed? I would expect so. This is another form of waste and something we want to eliminate. We also want to deliver value faster by automating. If it takes an entire day to release a new version of the app we’re developing before we start working on something new then that is waiting time. We can solve this via automation in multiple ways. Motion – Ruch zbędny Ever had a question but it requires that you gather people from across the organization? If so, this creates waste in that you have to have everyone move to a meeting room, participate in a meeting, and then eventually return to their workplace. What if we invited the wrong person which leads us astray? Another form of motion is around things like artifacts such as requirements. For example, if we have a requirement that changes a lot during the sprint then it requires a lot of potential re-work and confusion. This is unnecessary motion. Defects Defects in software or operations are the same as they are in manufacturing. It produces waste because we have to go back and fix the problem which prevents us from delivering new features. We can mitigate this by doing things like code reviews, automated code quality scans, following modern coding patterns, etc.
• #9: Corresponding to DORA with is DevOps Research and Assessment with was acquired now by google. DORA is creating free report named “State of DevOps Report” starring from 2014 year comparing over 31 thousands surveys. Comparing different values. And corresponding to report. Hight performer can do … 46 faster. Remember that: DevOps is the union of people, process, and products to enable continuous delivery of value to our end users So if we can reduce cycle we can deliver value faster.
• #10: So if we are able to unstart the cycle and act on it meaning continuous improve, to fast fail, fast learn, fast deliver we are to deliver value to customer and be top performer.
• #11: Some funny twit So we can end presentation ? No We need to remember that “DevOps is the union of people, process, and products to enable continuous delivery of value to our end users” And Azure DevOps is just tool and process that allow to do go this direction.
• #12: How to achieve this, by Azure DevOps
• #13: Lets first see history of Azure DevOps in time and see some related events in time. To have better context where we are now. There is Git. There Is DevOps There Is Open Source There Is Cloud. There Are Containers. There is Kubernetes or AKS there is Service Fabric or Fabric Mesh.
• #14: We show history and now let talk about Azure Boards
• #15: We have work items list to see items that are projects.
• #16: We can create /edit word items. We can also extend it to additional staff. Bu out from the box we have items that also to plan work and collaborate on them. So we have process that alow to work in it.
• #17: We have also board view so in easy way we can see what is going on and bringing transparency to team. All this can be extra customize if needed.
• #18: We have also work item backlog so we can prioritize, plan work. When we also work for longer time is we can try to forecast when we can deliver something and on this take decision on what and when deliver or not.
• #19: Depending from kind or work we can use sprint approach or kanban board.
• #20: We have option trace what is going on and we have analytics about status of projects.
• #21: What more about work items there are queries with alow to query work items.
• #22: There is option to crate or work with work items in bulk mode using excel.
• #23: Set switch now to talk about Azure Repos.
• #24: History.
• #25: We have branch policy so we can secure quality of code and process. Check for comment.
• #26: We can create pull request. So it is easy to trace what is changing. We can assignee work item, if we did not that this in commit.
• #27: So we create PR and branch policy is forcing us. To follow process and rules .
• #28: When we talk about git and code repository we need to talk about branch strategy. One best switch is covering 99% cases if Git Flow. And when we are following it is making it so much easier corresponding to not having standard approach how to and what to do when. Crate feature branch, hot fix. Release branch etc.
• #29: Let now talk about Azure Pipelines. What they are brining to us. …. Remember that we can build anything.
• #30: We create build definition by selecting source of code with can be Azure Repository, GitHub, GitHub Enterprise Server, Subversion ,Bitbucket or other git repository it can be also TFS Version Control. We chose project, repository and default branch.
• #31: Next we are select template of build and this can be anything that can run on Windows, Linux, Mac OS. We can create our own templates and share them with make sense if we are repeating the same pastern. We can also crate this as YAML definition but later about that.
• #32: When we chose some template or use empty one we are adding and configure build task. We are adding them to specific agent job. We have possibility to use multiple build agents at the same time or use different agent phases like agent less pahse. We can run job for multiple configurations to parallel it.
• #33: Here we are hosing the Agents pool on with we what to run job.
• #38: There variable group allow to integrate with Key vault. So we can easy integrate and keep secrets, certificate in one place.
• #39: So we schedule build and we are getting a lot of information. Like who trigger it, all logs tests that was run, linked work items types. We can see also history of runed build to see how it behave in time.
• #40: When we have unit tests in any technology stack and they have wrapper to generate output in specific format we are able to upload it and see test result and metrics.
• #41: There is build in analytic to see all the metric and also extension to use powerbi and crate custom reports.
• #42: YAML, I sad I will get back to it right now there is option to crate build definitions in GUI with are really a json under hood. But we have option to crate it as YAML. Definition and version it together with source code. But it is still in progress and work in going on to add more capabilities with are missing now.
• #43: Before we go next let put some words about cloud to understand time where we are now.
• #44: So we put some words about cloud and now let continue. Where we stopped, And what next that all no. Now we want to deploy our application. So we will copy it to our servers but we need to get budget hire multiple administrators, order servers, get electricity internet and backup for it, get licenses…, install systems, update systems. Open other location…. Backup policy….., use DSC OR PowerShell remoting to deploy it. What we wanted to do we wanted to do because I forgot. We wanted to deploy our application, and there is cloud, there is docker there is Infrastructure as code, there is docker.
• #45: So we care new git repository create project that is storing azure arm template with we also parametrize. We can use arm, azure cli, PowerShell, terraform but arm in best in that case that is described in way it will be later running under the hood, and it is independent. When comparing to others. So we describe what investiture we want and it will be setup this way, also there is way to validate it with out deploy. So we will create PaaS when we will have 2 azure function with are small web application connected to traffic manager witch is balancing traffic and also we will use application insight to monitor what is going on.
• #46: So we care release definition with can look like like this when we have multiple stages.
• #47: Each state looks very similar to build definition.
• #48: Each state looks very similar to build definition.
• #49: In Release pipeline we have also variable groups and key vault integration but we have also Scope to variable witch is connects to while release or to specific stage.
• #50: We have also artifacts for release definition with can be build definition, azure repository, GitHub, TFVC, Azure Artefacts, GitHub Release, Docker Hub, Jenkins. Also remember that inside each stage you can add task to for example download something from external source. Each artifacts cat trigger Release or we can setup time schedule. When we have multiple artefacts we can filtering for specific branch tag etc or use Gateawys.
• #51: Each stage can also have schedule. There is also pre deployment approval so specific person/group need to approve it.
• #52: Thee is artefact filter
• #53: Each state can have also Gats check to check external service of state and validate like azure policy, azure function, invoke rest api.
• #54: After Stage triggers when we orchestrate what stage is trigger after what stages
• #55: There are also post deployment for each stage.
• #56: Each release keep the history of used artefact so it easy to trace what when was release and what work items was assigned to it. So if can be used also to generate release notes.
• #57: Few word about build infrastructure.
• #60: There is Azure DevOps to is allowing to extend and crate custom extensions for Azure DevOps like GUI extension but also task build extensions.
• #61: Azure DevOps release by Azure DevOps
• #63: There are multiple types of tests the case is just to use them, smart.
• #64: We can crate test case from board from requirement
• #65: We can see that if is connected to User Story but this can by any other work item type
• #66: When we switch to test Plans We can see that test case “Welcome message ‘Name’” was create in Test Suite: 16:” And this test suit was create in as Test Plan (devOps_Team_Storeis_Iteration 1) with user current iteration and team name. So the level are Test Plans Test Suits Test Cases Crate test case is empty so let open it and add content.
• #69: So we can create suites for selected Requirements.
• #70: So no we created test suite from requirement and there is not test case , we have 3 option that we will add existing one, we will create new test case or we will crate test cases using grid.
• #71: So when we open grid view put the title and action or copy them from exel/csv we are able to build crate them. When we save tow new test cases will be crate and we can also see this on board of work item.
• #72: So we there is option to create parameters using @ and me test more usable passing multiple combinations, also there is option to create this as shared parameters so later on the set of parameters can be reused.
• #73: So we can select test cases or test plan containing test cases and we have few option to run. The is not also to use web browser testing, this is also used for desktop application testing. There can be more staff added like specific services working on stage servers to collect data from executing code that is running on server side. Usually how it is done, the test plan is prepared. And people are assigned to some test and they are doing test run. So by teas approach test that can not be automated or need to be run manual can be done by almost any one that will be introduced to this tool.
• #74: We have manual using web browser with is extension to Chrome or Fire Fox, Manual test using Test Runner Next one is Automated test using release stage with can we be can use build in test recording or we can use external one using like selenium and combine with attributes to integrate. Or can create fully programmatic test and decorate them to connect to test cases. https://docs.microsoft.com/en-us/azure/devops/test/run-automated-tests-from-test-hub?view=azure-devops 2 Last one are the Manual testing using old Test Plans client that was create from visual studio 2010 up to 2017 version, then there was no much web experiencers for Test Plans in web browser. Few years ago there was whole tool called Lab Center to setup environment by vms etc.
• #75: We have chose option to run with option so then we have option to assignee it with specific build.
• #76: So we have selecte test to run and we are doing this step by step of test case. There are few option we can capture screen/windows when need or we can crate screen capture for each step or we can event record video. At the end we can create bug from current windows or we can go to specific test run history in create bug from there.
• #78: So we have a vide or action log with is html with all screen that we taken and current test step. What is import
• #79: All kind of test can be wrapped and linked to work items to genera metrics with late on can be used. We can integrate with this tests from multiple lkangugae to have reports.
• #80: Any Packages Share efficient Package Caching for pipeline for hosted agent in preview In preview also public packages to share outside organization In can be used as symbol server so we are able having option of released version possibility to debug code.
• #81: We can manage mange Package When we crate Package our self and share across organization or projects we can manga with version in official release or prerelease We are able to secure build and deploy process by having it in out feed and do not be able to build when like few years ago npm was down for fee days. We can trace of usage for external and internal packages. What is important that Azure artefact can be used as symbol server.
• #82: How to crate feed
• #83: How to connect to feed
• #84: To use feed we can modify build restore step that will crate temporary nuget.config file, We select feed that what to use and unselect nuget.org depending from out setup.
• #85: And when running build we can see that packaged are added to feed because we use upstream source option for feed.
• #87: So when we will return what is DevOps so Azure DevOps is process, product, tool that is allowing to do this.
• #89: Thank You