Deploy an ASP.NET Core application to IIS using Octopus and GitHub Actions
In this tutorial, we show you how to build a fully-functional continuous delivery pipeline for a simple ASP.NET Core web application and deploy it to IIS. We use GitHub Actions to build the code and run tests, and we use Octopus Deploy to deploy and promote releases.
To get up and running quickly, the TestDrive VMs provide preconfigured environments demonstrating various continuous delivery pipelines documented in these guides.
Introduction
The application we'll deploy is called Random Quotes, which is a simple web application that randomly displays a famous quote each time the page loads. It consists of a web front end and a database that contains the quotes. We'll build a complete Continuous Integration/Continuous Delivery (CI/CD) pipeline with automated builds, deployments to a dev environment, and sign offs for production deployments.
Deployment pipeline
For this tutorial, we assume you use Git for version controlling changes to your source code and GitHub Actions to compile code and run unit tests. Octopus Deploy will take care of the deployment. Here is what the full continuous integration and delivery pipeline will look like when we are finished:
The development team's workflow is:
- Developers commit code changes to Git.
- GitHub Actions detects the change and performs the continuous integration build, this includes resolving any dependencies and running unit tests.
- When the GitHub Actions build completes, the change will be deployed to the Dev environment.
- When one of your team members (perhaps a tester) wants to see what's in a particular release, they can use Octopus to manually deploy a release to the Test environment.
- When the team is satisfied with the quality of the release and they are ready for it to go to production, they use Octopus to promote the release from the Test environment to the Production environment.
Since Octopus is designed to be used by teams, in this tutorial we also set up some simple rules:
- Anyone can deploy to the dev or test environments.
- Only specific people can deploy to production.
- Production deployments require sign off from someone in our project stakeholders group.
- We'll send an email to the team after any test or production deployment has succeeded or failed.
This tutorial makes use of the following tools:
Version control of source code
Compile code, run tests
Release management & deployments
Hosts the running web application
Octopus is an extremely powerful deployment automation tool, and there are numerous ways to model a development team's workflow in Octopus Deploy, but this tends to be the most common for small teams. If you're not sure how to configure Octopus, we recommend following this guide to learn the basics. You'll then know how to adjust Octopus to suit your team's workflow.
This tutorial takes about an hour to complete. That sounds like a long time, but keep in mind, at the end of the tutorial, you'll have a fully-functional CI/CD environment for your entire team, and you'll be ready to deploy to production at the click of a button. It's worth the effort!
Build vs. deployment
For any non-trivial application, you're going to deploy the software to multiple environments. For this tutorial, we're using the environments Dev, Test, and Prod. This means you need to choose between building your application once or building it before each deployment? To reduce the risk of a failed production deployment, Octopus strongly encourages the practice of building once, and deploying multiple times.
The following activities are a build time concern, so they will happen in GitHub Actions after any change to code is committed to Git:
- Check out the latest changes from Git.
- Resolve and install any dependencies from NuGet.
- Run unit tests.
- Package the application by bundling all the files it needs to run into a ZIP file.
This results in a green CI build and a file that contains the application and everything it needs to run. Any configuration files will have their default values, but they won't know anything about dev vs. production settings just yet.
Lastly, it's very important that we give this artifact a unique version number. We will produce a new artifact file every time our CI build runs, and we don't want to accidentally deploy a previous version of the artifact.
An example of a package that is ready to be deployed is:
RandomQuotes.1.0.0.zip
At this point, we have a single artifact that contains all the files our application needs to run, ready to be deployed. We can deploy it over and over, using the same artifact in each environment. If we deploy a bad release, we can go and find the older version of the artifact and re-deploy it.
The following activities happen at deployment time by Octopus Deploy:
- Changing any configuration files to include settings appropriate for the environment, e.g., database connection strings, API keys, etc.
- Uploading the artifacts to IIS.
- Running tasks that need to be performed during the deployment such as database migrations or taking the application temporarily offline.
Prerequisites
There are a number of tools you need to install to implement a complete CI/CD workflow. These include the GitHub Actions and Octopus servers, some command-line tools, and IIS to host the final deployment.
Git
The source code for the sample application is hosted on GitHub. To access the code, you need the Git client. The Git documentation has instructions to download and install the Git client.
GitHub Actions
GitHub Actions are available for free on public GitHub repsoitories and with paid options for private repositories.
.NET Core SDK
The .NET Core SDK version 2.2 or above is required to compile and run the ASP.NET Core sample application. The SDK can be downloaded from the Microsoft .NET website.
The .NET Core SDK must be installed locally for development and also on the GitHub Actions server or any agents that will compile the application code.
.NET Core Hosting Bundle
Hosting .NET Core applications in IIS requires the .NET Core Hosting Bundle to be installed.
Getting Started with Octopus Cloud
Before you can start an Octopus Cloud trial, you'll need an Octopus account.
You can sign up for an account at: octopus.com/register.
Create an Octopus Account
An Octopus account lets you manage your instances of Octopus Cloud.
- Enter your name.
- Provide your email address and click Create a password. Please note, these credentials are for your Octopus Account. You will also create credentials for your Octopus Cloud instance, when you create it.
- On the next screen, provide your company name.
- Chose a secure password and enter it twice.
- Click Create my Octopus account.
Create a Cloud Instance
- From the instances screen, click Create cloud instance.
- Enter an instance name for your Octopus Cloud instance.
- Choose a URL for the instance.
- Select the Cloud region for your instance. Currently the only option is US - Oregon.
- Click Enter account details.
- Create your first user for Octopus Cloud.
- Enter the username the user will use to log into Octopus Cloud.
- Create a password for the user and confirm the password.
- Click Continue to Confirmation.
- Confirm the details you've provided, agree to the terms, and click Looks good. Deploy my Octopus!.
You will be taken to the account provisioning screen. Please note it can take five to ten minutes for your Octopus Cloud instance to be ready. You will receive an email when the instance is ready to use.
When the instance is ready, you will see it (and any other instances you have access to) the next time you log in to your Octopus account at https://account.octopus.com/account/signin.
IIS
Microsoft IIS is bundled with Microsoft Windows. See the Microsoft IIS website for instructions on how to install IIS in Windows.
The following command will enable the required IIS components in Windows:
Start /w pkgmgr /iu:IIS-WebServerRole;IIS-WebServer;IIS-CommonHttpFeatures;IIS-StaticContent;IIS-DefaultDocument;IIS-DirectoryBrowsing;IIS-HttpErrors;IIS-ApplicationDevelopment;IIS-ASPNET;IIS-NetFxExtensibility;IIS-ISAPIExtensions;IIS-ISAPIFilter;IIS-HealthAndDiagnostics;IIS-HttpLogging;IIS-LoggingLibraries;IIS-RequestMonitor;IIS-Security;IIS-RequestFiltering;IIS-HttpCompressionStatic;IIS-WebServerManagementTools;IIS-ManagementConsole;WAS-WindowsActivationService;WAS-ProcessModel;WAS-NetFxEnvironment;WAS-ConfigurationAPI;IIS-ASPNET45
Clone source code
The source code for the random quotes application is hosted in GitHub. The code can be cloned from https://github.com/OctopusSamples/RandomQuotes.git with the command:
git clone https://github.com/OctopusSamples/RandomQuotes.git
The RandomQuotes application can be run locally with the command:
cd RandomQuotes/RandomQuotes dotnet run
Octopus API key
In order to allow GitHub Actions to communicate with Octopus, we need to generate an API key. This is done in the Octopus web portal. The web portal can be opened from the Browse link in the Octopus Manager:
From the Octopus Deploy web portal, sign in, and view your profile:
Go to the API keys tab. This lists any previous API keys that you have created. Click on New API key:
Give the API key a name so that you remember what the key is for, and click Generate New:
Copy the new API key to your clipboard:
Continuous integration
GitHub Actions supports Node.js, Python, Java, Ruby, PHP, Go, Rust, .NET, and more. Build, test, and deploy applications in your language of choice. In this tutorial, we rely on GitHub Actions to do the following:
- Clone the code from Git.
- Resolve and install any dependencies from NuGet.
- Run unit tests.
- Package the application by bundling all the files it needs to run into a ZIP file.
- Push the package to the Octopus (built-in) package repository.
At a high level, GitHub Actions workflows are YAML documents committed to a GitHub repsository that define the runner that the steps are executed on, the triggers that execute the workflow, and the steps to be run.
GitHub Actions workflows have access to encrypted secrets. Two secrets are defined called:
- OCTOPUS_SERVER_URL, which defines the URL of the Octopus server (for example, https://myserver.octopus.app for hosted Octopus instances)
- OCTOPUS_API_TOKEN, which defines the API key used to access the Octopus instance.
The YAML document below is saved to a file called .github/workflows/build.yaml. It is configured to be triggered on a git push and defines steps to build, test, package, and publish the sample application:
jobs: build: runs-on: ubuntu-latest steps: - uses: actions/checkout@v3 with: fetch-depth: '0' - name: Set up DotNET Core uses: actions/setup-dotnet@v3 with: dotnet-version: 3.1.402 - name: Install GitVersion uses: gittools/actions/gitversion/setup@v0.9.14 with: versionSpec: 5.x - id: determine_version name: Determine Version uses: gittools/actions/gitversion/execute@v0.9.14 with: additionalArguments: /overrideconfig mode=Mainline - name: Install Octopus Deploy CLI uses: OctopusDeploy/install-octopus-cli-action@v1 with: version: latest - name: Install Dependencies run: dotnet restore shell: bash - name: Test run: dotnet test -l:trx shell: bash - if: always() name: Report uses: dorny/test-reporter@v1 with: name: DotNET Tests path: '**/*.trx' reporter: dotnet-trx fail-on-error: 'false' - name: Publish run: dotnet publish --configuration Release /p:AssemblyVersion=${{ steps.determine_version.outputs.assemblySemVer }} - id: package name: Package run: | # Find the publish directories shopt -s globstar paths=() for i in **/publish/*.dll; do dir=${i%/*} echo ${dir} paths=(${paths[@]} ${dir}) done eval uniquepaths=($(printf "%s\n" "${paths[@]}" | sort -u)) for i in "${uniquepaths[@]}"; do echo $i done # For each publish dir, create a package packages=() versions=() for path in "${uniquepaths[@]}"; do # Get the directory name four deep, which is typically the project folder. # The directory name is used to name the package. dir=${path}/../../../.. parentdir=$(builtin cd $dir; pwd) projectname=${parentdir##*/} # Package the published files octo pack \ --basePath ${path} \ --id ${projectname} \ --version ${{ steps.determine_version.outputs.semVer }} \ --format zip \ --overwrite packages=(${packages[@]} "${projectname}.${{ steps.determine_version.outputs.semVer }}.zip") versions=(${versions[@]} "${projectname}:${{ steps.determine_version.outputs.semVer }}") done # Join the array with commas printf -v joined "%s," "${packages[@]}" # Save the list of packages as an output variable echo "::set-output name=artifacts::${joined%,}" # Do the same again, but use new lines as the separator # https://trstringer.com/github-actions-multiline-strings/ # Multiline strings require some care in a workflow printf -v versionsjoinednewline "%s\n" "${versions[@]}" versionsjoinednewline="${versionsjoinednewline//'%'/'%25'}" versionsjoinednewline="${versionsjoinednewline//$'\n'/'%0A'}" versionsjoinednewline="${versionsjoinednewline//$'\r'/'%0D'}" # Save the list of packages newline separated as an output variable echo "::set-output name=versions_new_line::${versionsjoinednewline%\n}" - name: Push packages to Octopus Deploy 🐙 uses: OctopusDeploy/push-package-action@v2 env: OCTOPUS_API_KEY: ${{ secrets.OCTOPUS_API_TOKEN }} OCTOPUS_HOST: ${{ secrets.OCTOPUS_SERVER_URL }} with: overwrite_mode: OverwriteExisting packages: ${{ steps.package.outputs.artifacts }} name: DotNET Core Build 'on': workflow_dispatch: {} push: {}
There is a lot of work being performed by this workflow, so let's break it down.
The job called build is run on an Ubuntu runner:
jobs: build: runs-on: ubuntu-latest
The first step checks out the source code from the git repository. Setting the fetch-depth option to 0 means all the history for all branches and tags is fetched. This is required by the GitVersion action used in later steps:
steps: - uses: actions/checkout@v3 with: fetch-depth: '0'
This step sets up DotNET Core 3 which is used to build the application source code:
- name: Set up DotNET Core uses: actions/setup-dotnet@v3 with: dotnet-version: 3.1.402
GitVersion builds meaningful version strings based on the commits to git.
This step installs GitVersion:
- name: Install GitVersion uses: gittools/actions/gitversion/setup@v0.9.14 with: versionSpec: 5.x
This step calls GitVersion which outputs many different variables representing different parts of a SemVer version string generated from the commits to the git repository. Later steps use these output values as the version of the package generated by this workflow:
- id: determine_version name: Determine Version uses: gittools/actions/gitversion/execute@v0.9.14 with: additionalArguments: /overrideconfig mode=Mainline
This step installs the Octopus CLI, which is used by the subsequent Octopus GitHub Actions:
- name: Install Octopus Deploy CLI uses: OctopusDeploy/install-octopus-cli-action@v1 with: version: latest
This step installs the application's dependencies:
- name: Install Dependencies run: dotnet restore shell: bash
This step runs the unit tests:
- name: Test run: dotnet test -l:trx shell: bash
This step publishes the results of the test against the workflow run. The if parameter is set to always() to ensure that the results are published even if the tests fail:
- if: always() name: Report uses: dorny/test-reporter@v1 with: name: DotNET Tests path: '**/*.trx' reporter: dotnet-trx fail-on-error: 'false'
This step publishes the application's executable files. The version of the assemblies is set to the assemblySemVer output generated by GitVersion:
- name: Publish run: dotnet publish --configuration Release /p:AssemblyVersion=${{ steps.determine_version.outputs.assemblySemVer }}
This step packages the application's published executable files into a self contained artifact:
- id: package name: Package run: |
The bash script defined in this step relies on the conventions used by the dotnet publish command to determine the location of the published executable files.
By default, the dotnet publish command saves files to the directory [project_file_folder]/bin/[configuration]/[framework]/publish/ .
This kind of generalized logic is perhaps overkill for the Random Quotes sample application, because there is only one project being built and we know exactly where the published files will be found. However, more complex DotNET applications can have several projects referenced by a single solution, in which case multiple projects are built and published. The logic presented here accomodates these more complex projects.
The first part of the bash script finds dll files located under a directory called publish and saves the directory to a variable called paths:
# Find the publish directories shopt -s globstar paths=() for i in **/publish/*.dll; do dir=${i%/*} echo ${dir} paths=(${paths[@]} ${dir}) done
There are likely many dll files, meaning the paths variable contains many duplicate entries. The variable uniquepaths is created containing the unique values found in the paths variable:
eval uniquepaths=($(printf "%s\n" "${paths[@]}" | sort -u))
Each of the unique paths is printed to the console for debugging:
for i in "${uniquepaths[@]}"; do echo $i done
We then define two arrays that hold the paths to packages artifacts and the versions:
packages=() versions=()
We loop over the directories containing the published files:
for path in "${uniquepaths[@]}"; do
The package ID is determined by the directory containing the project file. Remember that the packaged files are saved to the directory [project_file_folder]/bin/[configuration]/[framework]/publish/ by default. This means the project directory is found four levels up from the published files in the directory structure:
# Get the directory name four deep, which is typically the project folder. # The directory name is used to name the package. dir=${path}/../../../.. parentdir=$(builtin cd $dir; pwd) projectname=${parentdir##*/}
The published files are packaged as a Zip archive with the ID determined from the project directory name and with the version contained in the output variable semVer generated by GitVersion:
# Package the published files octo pack \ --basePath ${path} \ --id ${projectname} \ --version ${{ steps.determine_version.outputs.semVer }} \ --format zip \ --overwrite
The location of the zip file is added to the packages variable, and the version of the package is added to the versions variable:
packages=(${packages[@]} "${projectname}.${{ steps.determine_version.outputs.semVer }}.zip") versions=(${versions[@]} "${projectname}:${{ steps.determine_version.outputs.semVer }}")
Actions use lists in a few different ways. Some actions require lists to be comma separated, while others use newline separated lists.
The list of packages is exported a comma separated list called artifacts:
# Join the array with commas printf -v joined "%s," "${packages[@]}" # Save the list of packages as an output variable echo "::set-output name=artifacts::${joined%,}"
The list of package versions is exported as a newline separated list called versions_new_line.
In order to create a newline separated list that can be consumed in a later step, we must take care to escape special characters like the percent sign and line breaks:
# Do the same again, but use new lines as the separator # https://trstringer.com/github-actions-multiline-strings/ # Multiline strings require some care in a workflow printf -v versionsjoinednewline "%s\n" "${versions[@]}" versionsjoinednewline="${versionsjoinednewline//'%'/'%25'}" versionsjoinednewline="${versionsjoinednewline//$'\n'/'%0A'}" versionsjoinednewline="${versionsjoinednewline//$'\r'/'%0D'}" # Save the list of packages newline separated as an output variable echo "::set-output name=versions_new_line::${versionsjoinednewline%\n}"
The package files are then pushed to Octopus:
- name: Push packages to Octopus Deploy uses: OctopusDeploy/push-package-action@v2 env: OCTOPUS_API_KEY: ${{ secrets.OCTOPUS_API_TOKEN }} OCTOPUS_HOST: ${{ secrets.OCTOPUS_SERVER_URL }} with: overwrite_mode: OverwriteExisting packages: ${{ steps.package.outputs.artifacts }}
The workflow is triggered on each push to the git repository. The workflow can also be triggered manually with the workflow_dispatch event:
name: DotNET Core Build 'on': workflow_dispatch: {} push: {}
Commit the file and push it to the GitHub repository with the command:
git add .; git commit -m "Added workflow file"; git push
GitHub Actions then runs the workflow.
Deploying with Octopus Deploy
Now that GitHub Actions has successfully built the application, we need to configure Octopus to deploy it into our environments.
Create the environments
Environments represent the stages that a deployment must move through as part of the deployment pipeline. We'll create three environments: Dev, Test, and Prod.
Log into Octopus, and click the Infrastructure link, then the Environments link, and click ADD ENVIRONMENT:
Enter Dev as the New environment name, and click SAVE:
Repeat the process for the Test and Prod environments:
Deployment Targets
In order to deploy the application, we need to configure a target. For this tutorial, we will deploy the ASP.NET Core applications via a polling Tentacle.
Tentacle Manager is the Windows application that configures your Tentacle. The links below provide downloads for the Tentacle Windows installer.
Once installed, you can access it from your start menu/start screen. Tentacle Manager can configure Tentacles to use a proxy, delete the Tentacle, and show diagnostic information about the Tentacle.
- Start the Tentacle installer, accept the license agreement, and follow the onscreen prompts.
- When the Octopus Deploy Tentacle Setup Wizard has completed, click Finish to exit the wizard.
- When the Tentacle Manager launches, click GET STARTED.
- On the communication style screen, select Polling Tentacle and click Next.
- If you are using a proxy see Proxy Support, or click next.
-
Add the Octopus credentials the Tentacle will use to connect to the Octopus Server:
- The Octopus URL: the hostname or IP address.
-
Select the authentication mode and enter the details:
- The username and password you use to log into Octopus, or:
- The Octopus API key created earlier.
- Click Verify credentials, and then next.
-
Give the machine a meaningful name and assign the Tentacle to all the environments (Dev, Test, and Prod) we created earlier.
For this tutorial we will use a single physical target, and therefore a single Tentacle, to host all our environments.
- Choose or create the role web.
- Leave Tenants and Tenant tags blank.
- Click Install, and when the script has finished, click Finish.
Create the Octopus deployment project
With the environments defined and a target created, we now need to create a deployment project in Octopus.
Log into Octopus, click the Projects link, and click ADD PROJECT:
Enter Random Quotes for the project name, and click SAVE:
We start by defining the variables that we will consume as part of the deployment. Click the Variables link, and then select the Project option:
The first variable we'll define is called AppSettings:EnvironmentName. This matches the name of a key in the appsettings.json file:
{ ... "AppSettings": { ... "EnvironmentName": "DEV" } }We set the value of this variable to match the name of the environment that Octopus is deploying the application into by setting its value to #{Octopus.Environment.Name}. This variable, in conjunction with the Structured Configuration Variables feature, means that the config file will be updated during deployment to reflect the current environment. This is how we can have a single, environment agnostic package, and customize it for any environment it is deployed into.
We are then going to define a variable called IIS Port with unique values bound to each environment. This means that when we deploy the project to each environment, a new port will be used to expose the application in IIS.
By exposing each environment as a unique port, we can implement multiple environments with a single physical target.
- Define a value of 8081 scoped to the Dev environment.
- Define a value of 8082 scoped to the Test environment.
- Define a value of 8083 scoped to the Prod environment.
Then click SAVE to save the changes:
We will now define the deployment process. Click Deployments link, then the Overview link, and then click DEFINE YOUR DEPLOYMENT PROCESS:
Click ADD STEP:
Enter iis into the search box:
Click ADD on the IIS tile:
We want to take advantage of a feature in Octopus that replaces settings in the web.config file based on our variables. Specifically, we're replacing the AppSettings:EnvironmentName variable that was defined earlier. To do this we enable the Structured Configuration Variables feature.
Click CONFIGURATION FEATURES:
Enable the Structured Configuration Variables option, and click OK:
In the Structured Configuration Variables section, enter appsettings.json as the Target files value.
Enter Deploy web app to IIS as the step name:
Select the web role. This role matches the role assigned to the Tentacle that was configured earlier.
Select the RandomQuotes package ID. This is the package that GitHub Actions pushed to the built-in feed:
Enter RandomQuotes as the web site name:
Enter RandomQuotes as the application pool name:
Because we have installed Octopus and IIS on the same host, and Octopus is listening on port 80, we need to remove the default port binding to avoid a conflict. Click the cross icon on the existing binding:
Click ADD to add a new binding:
Enter #{IIS Port} as the port. This is how we reference the IIS Port variable created earlier. Then click OK:
Select the Enable Anonymous authentication option, and unselect the Enable Windows authentication option. Then click SAVE:
Deploy!
We now have a deployment project in Octopus ready to deploy our ASP.NET Core application to our Dev, Test, and Prod environments. The next step is to create and deploy a release.
Click CREATE RELEASE.
The release creation screen provides an opportunity to review the packages that will be included and to add any release notes. By default, the latest package is selected automatically. Click SAVE:
This screen allows you to select the environment that will be deployed into. Lifecycles can be used to customize the progression of deployments through environments (this is demonstrated later in the guide), however, we will accept the default option to deploy to the Dev environment by clicking DEPLOY TO DEV...:
Click DEPLOY to deploy the application into the Dev environment:
The application is then deployed:
Congratulations! You have now built and deployed your first application. Visit http://localhost:8081 in a browser to view a random quote. Note the port (8081) matches the value of the IIS Port variable we scoped to the Dev environment.
Continuous deployments
The process of deploying a successful build to the Dev environment is currently a manual one; GitHub Actions pushes the file to Octopus, and we must trigger the initial deployment to the Dev environment from within Octopus. Typically though, deployments to the Dev environment will be performed automatically if a build and all of its tests succeed.
To trigger the initial deployment to the Dev environment after a successful build, we will go back to the project we created in GitHub Actions and add an additional step to create an Octopus release and then deploy it to the Dev environment.
When added to the workflow file, the YAML below creates a new release in Octopus with each build and deploys it to the development environment.
Note that the package versions are explicitly defined via the output variable called versions_new_line generated in previous steps:
- name: Create Octopus Release uses: OctopusDeploy/create-release-action@v1.1.1 with: api_key: ${{ secrets.OCTOPUS_API_TOKEN }} project: Random Quotes server: ${{ secrets.OCTOPUS_SERVER_URL }} deploy_to: Dev packages: ${{ steps.package.outputs.versions_new_line }}
Additional configuration
Now we will explore some of the more advanced features of Octopus that allow us to customize the deployment progression through environments, secure deployments to production environments, add deployment sign offs, view the audit logs, and add notifications.
Lifecycles
Our project currently uses the default lifecycle, which defines a progression through all the environments in the order they were created.
A custom lifecycle allows the progression of a deployment to be further refined, by defining only a subset of environments that can be deployed to, allowing some environments to be skipped entirely, or requiring that a minimum number of environments are successfully deployed to before moving onto the next environment.
Here we will create a custom lifecycle that makes deployments to the Dev environment optional. This means that initial deployments can be made to the Dev or Test environments, but a successful deployment must be made to the Test environment before it can be progressed to the Prod environment.
Skipping the Dev environment like this may be useful for promoting a release candidate build directly to the Test environment for product owners or testers to review.
Click the Library link, click the Lifecycles link, and click ADD LIFECYCLE:
Set the lifecycle name to Dev, Test, and Prod, and the description to Progression from the Dev to the Prod environments:
Phases are used to group environments that can accept a deployment. Simple lifecycles, such as the lifecycle we are creating, have a one-to-one relationship between phases and environments.
Click ADD PHASE:
Enter Dev as the phase name, and select the Optional phase option. This means that deployments can skip this phase and any environments defined in it, and deploy directly to the next phase.
Because we are mapping each environment to its own phase, the name of the phase matches the name of the environment:
Click ADD ENVIRONMENT:
Click the dropdown arrow, select the Dev environment, and click OK:
Repeat the process to add a new phase for the Test and Prod environments, leaving the default All must complete option selected:
Click SAVE:
Now, we need to switch the deployment project from the Default Lifecycle to the newly created lifecycle.
Click the Projects link, and click the Random Quotes project tile:
Click the Process link, and click CHANGE:
Select the Dev, Test, and Prod lifecycle. Notice the Dev environment is shown as optional.
Click SAVE:
Click CREATE RELEASE, and click SAVE to save the new release:
Because the Dev environment has been configured as optional, the initial release can be made to either the Dev or Test environments. We'll skip the Dev environment and deploy straight to Test.
Click DEPLOY TO..., and select the Test environment:
Click DEPLOY to deploy the application to the Test environment:
The deployment is then performed directly in the Test environment, skipping the Dev environment:
Opening http://localhost:8082 displays the copy of the Random Quotes application deployed to the Test environment. Note the port 8082 is the value of the IIS Port variable scoped to the Test environment.
Also, see the footer text that says running in Test. This is the result of the Configuration variables feature and the EnvironmentName variable replacing the initial value of DEV in the Web.config file with Test. This is an example of a single, environment agnostic package being customized for the environment it is being deployed into:
Approvals
It's a common business requirement to have testers or product owners manually verify that a particular build meets the requirements before a deployment can be considered successful.
Octopus supports this workflow through the use of manual intervention steps. We'll add a manual intervention step to the end of the deployment process, which requires a responsible party to verify the build meets the requirements.
Open the Random Quotes project, click the Process link, and click the ADD STEP button:
Search for the Manual Intervention Required step, and add it to the process:
Enter Deployment Sign Off for the Step Name:
Enter the following for the Instructions:
Open the application and confirm it meets all the requirements.
Because every build is automatically deployed to the Dev environment, it doesn't make sense to force someone to manually approve all those deployments. To accommodate this, we do not enable the manual intervention step for deployments to the Dev environment.
Expand the Environments section under the Conditions heading, select the Skip specific environments option, and select the Dev environment.
Click SAVE to save the step:
When this application is deployed to the Test or Prod environments, a prompt will be displayed requiring manual sign off. Click ASSIGN TO ME to assign the task to yourself:
Add a note in the provided text box, and click PROCEED to complete the deployment:
The deployment will then complete successfully:
Email notifications
Octopus has native support for sending email notifications as part of the deployment process. We will add a step to the deployment process to let people know when a release has been deployed to an environment.
To start, we need to configure an SMTP server to send our emails. For this guide, we'll use the free SMTP server provided by Google.
Click the Configuration link:
Click the SMTP link:
- Enter smtp.gmail.com as the SMTP Host.
- Enter 587 as the SMTP Port.
- Enable the Use SSL/TLS option.
- Enter your Gmail address as the From Address.
- Enter your Gmail address and password in the credentials.
You will enable the Less secure apps option on your Google account for Octopus to send emails via the Google SMTP server.
Open the Random Quotes project, click the Process link, and click ADD STEP:
Search for the Send an Email step, and add it to the process:
Enter Random quotes deployment status for the Step Name:
Enter the email address to receive the notification in the To field:
Enter Random quotes deployment status as the Subject:
Enter the following as the Body:
Deployment to #{Octopus.Environment.Name} #{each step in Octopus.Step} StepName: #{step} Status: #{step.Status.Code} #{/each}
Here we use the #{Octopus.Environment.Name} variable provided by Octopus to add the name of the environment that was deployed to, and then loop over the status codes in the #{Octopus.Step} variable to return the status of each individual step.
The complete list of system variables can be found in the Octopus documentation.
We want to be notified of the status of this deployment regardless of whether the deployment succeeded or failed. Click the Run Conditions section to expand it.
Select the Always run option, which ensures the notification email is sent even when the deployment or the manual intervention fail:
Given every change to the source code will result in a deployment to the Dev environment, we do not want to generate reports for deployments to this environment.
Click the Environments section to expand it. Select the Skip specific environments option, and select the Dev environment to skip it.
This is the last piece of configuration for the step, so click SAVE to save the changes:
Deploy the project to the Test or Prod environments. When the deployment succeeds, the notification email will be sent:
Permissions
One of the strengths of Octopus is that it models environments as first-class entities. This means the security layer can apply rules granting access only to specific environments. We'll take advantage of this ability to model and secure environments to create two users, an internal deployer who can deploy to the Dev and Test environments, and a production deployer who can deploy to the Prod environment.
We start by creating the users. Click the Configuration link:
Click the Users link:
Click ADD USER:
Enter internaldeployer as the Username:
Enter Internal Deployer as the Display Name:
Enter the user's email address. We have used a dummy address of internaldeployer@example.org here:
Enter a password and confirm it. Then click SAVE to create the user:
Repeat the process for a user called productiondeployer. The summary for the productiondeployer user is shown below:
The newly created users are not assigned to any teams and have no permissions to do anything. To grant them permissions, we must first create two teams. The internal deployment team will grant access to deploy to the Dev and Test environments, while the production deployment team will grant access to deploy to the Prod environment.
Click the Configuration link:
Click the Teams link:
Click ADD TEAM:
Enter Internal Deployers for the New team name, and Grants access to perform a deployment to the internal environments for the Team description. Click SAVE to create the team:
We need to add the internaldeployer user to this team. Click ADD MEMBER:
Select the Internal Deployer user from the dropdown list, and click ADD:
The team does not grant any permissions yet. To add permissions click the USER ROLES tab, and click INCLUDE USER ROLE:
Select the Deployment creator role from the dropdown list. As the name suggests, this role allows a user to create a deployment, which results in the deployment process being executed.
Click DEFINE SCOPE:
We only want to allow the internal deployer to deploy to the internal environments. Select the Dev and Test environments, and click APPLY:
The permissions are then applied. We need a second permission to allow the internal deployer to view the projects dashboard. Click INCLUDE USER ROLE again:
Select the Project viewer role. This role does not need to be scoped, so click the APPLY button:
Here are the final set of roles applied to the team:
Repeat the process to create a team called Production Deployers that includes the productiondeployer user, and grants the Deployment creator role scoped to the Prod environment:
When we log in as the internaldeployer user, we see that the Random Quotes project dashboard shows DEPLOY... buttons for the Dev and Test environments. Any deployments in the production environment will be visible, but they cannot be created by this user:
When we log in as the productiondeployer user, we see that the Random Quotes project dashboard shows DEPLOY... buttons only for the Prod environment. Also note that the lifecycle rules still apply, and only successful deployments to the Test environment are available to be promoted to the Prod environment:
Audit Log
Important interactions in Octopus are tracked in the audit log. This is useful for teams that have security or legislative requirements to track the changes and deployments made to their infrastructure.
To view the audit log, click the Configuration link:
Click the Audit link:
A complete list of records are shown, with filtering available to help find specific events:
Conclusion
In this guide we ran through the process of building a complete CI/CD pipeline with:
- GitHub Actions building and testing the source code and pushing the package to Octopus.
- Octopus deploying the package to the Dev, Test, and Prod environments.
- Email notifications generated when deployments succeed or fail.
- Manual sign off for deployments to the Test and Prod environments.
- Users with limited access to create releases in a subset of environments.
This is a solid starting point for any development team, but Octopus offers so much more! Below are more resources you can access to learn about the advanced functionality provided by Octopus:
- The Octopus documentation
- The Octopus blog
- Upcoming events and webinars
- The Octopus community slack channel
Need support? We're here to help.