Redeploying

The Pittsburgh Purchasing Suite is built primarily on top of three key services: Heroku, AWS S3, and Sendgrid (though the email delivery service is fairly agnostic, as will be discussed later). Building out a production-level version of the Pittsburgh Purchasing Suite involves configuring these services and setting up some key environment variables via the Heroku UI.

Provisioning Services

Heroku

Heroku is a cloud platform based on a managed container system, with integrated data services and a powerful ecosystem, for deploying and running modern apps. The Heroku developer experience is an app-centric approach for software delivery, integrated with today’s most popular developer tools and workflows.

Heroku is a Platform as a Service (PaaS) that simplifies deployment and production server management. The Pittsburgh Purchasing Suite runs on top of Heroku. It uses web processes to handle web requests, worker processes to handle sending emails (see Notification), rebuilding the search index (see RefreshSearchViewMixin), and other background tasks. It also uses the Heroku scheduler add-on to do nightly jobs (see Nightly Jobs).

The base Heroku requirements involve Heroku Postgres and Heroku Redis. Redeploying a version as it currently exists requires both of these as dependencies.

AWS S3

Amazon Simple Storage Service (Amazon S3), provides developers and IT teams with secure, durable, highly-scalable object storage. Amazon S3 is easy to use, with a simple web service interface to store and retrieve any amount of data from anywhere on the web.

S3 is a storage services used to store PDF documents, specifically .pdf versions of the contracts themselves, along with documents that exist to support opportunities. S3 also stores all static assets (img/js/css) for the application. By default, uploaded documents have aggressive caching headers and the maximum expiry set dates.

Sendgrid

SendGrid is a cloud-based SMTP provider that allows you to send email without having to maintain email servers. SendGrid manages all of the technical details, from scaling the infrastructure to ISP outreach and reputation monitoring to whitelist services and real time analytics.

Sendgrid is an email provider that gives hooks for web-based and SMTP-based email. The suite uses Flask Mail. Because of this, you should be able to swap out any similar email provider (such as Mandrill, etc.) and get similar functionality as long as you update the MAIL_USERNAME and MAIL_PASSWORD environment variables.

Configuring the Environment

In order to deploy the Pittsburgh Purchasing Suite, the following environmental variables must be set via the Heroku settings page. For more information, see this heroku guide.

• SECRET_KEY

  The secret key protects the app’s session, forms, and other pieces. For more information about how Flask uses secret keys, please look at this stack overflow answer. To generate a secret key, you can run the following snippet directly from bash:

  python -c "import os; print os.urandom(24).encode('hex')"
  

  Take that and input it as the value to the SECRET_KEY

• BROWSERID_URL

  The app uses persona to manage authentication (see User Logins for more). BROWSERID_URL is domain to send over to Persona to validate that people are logging into the right machine. For local development, this will be 127.0.0.1:9000, and for production, it should match whatever users see in their address bar.

• MAIL_DEFAULT_SENDER

  The default from email address that will be used for sending out emails – if nothing is provided from the environment, this will default to to no-reply@buildpgh.com.

• BEACON_SENDER

  The default from email address that will be used for sending out emails through Beacon. If nothing is provided from the environment, this will default to beaconbot@buildpgh.com.

• CONDUCTOR_SENDER

  The default from email address that will be used for sending out emails through Beacon. If nothing is provided from the environment, this will default to conductorbot@buildpgh.com.

• MAX_CONTENT_LENGTH

  The maximum file size for files to be uploaded. This defaults to 2MB.

• UPLOAD_FOLDER

  The upload directory is the temporary home for uploaded conductor COSTARS files (see Uploads).

• S3_BUCKET_NAME

  The bucket name for Amazon S3 files (which will hold static assets like CSS/JS). No default.

• AWS_ACCESS_KEY_ID

  Your username for your AWS account. For more information, see AWS IAM documentation

• AWS_SECRET_ACCESS_KEY

  Your password for your AWS account. For more information, see AWS IAM documentation

• SERVER_NAME

  The name and port number of the server. For more, see the Flask documentation on SERVER_NAME.

• DISPLAY_TIMEZONE

  A timezone string that will be used for display times across the application. This must be a valid pytz string, which is a highly exhaustive list. For more, see the pytz docs.

• EXTERNAL_LINK_WARNING

  For dotgov.gov compliance, links to external sites must be flagged with a small alert letting people know they are leaving the site. This is a boolean flag as to whether this should be enabled or disabled. Defaults to False.

• DATABASE_URL

  The link to the database. If Herkou is being used, this value will be set automatically.

• MAIL_USERNAME

  The username to the mail send service being used. If using a service like Sendgrid, this can be configured as a separate custom account from the accounts dashboard.

• MAIL_PASSWORD

  The password to the mail send service being used.

• CELERY_BROKER_URL

  The URL to the celery broker that will handle the offline jobs. Defaults to REDIS_URL, a value provided by Heroku redis. Other brokers can be used (such as, for example, RabbitMQ or others). For more information about that, take a look at the celery documentation

• CELERY_RESULT_BACKEND

  The location of where to store the results from any celery long-running tasks.

• CACHE_REDIS_URL

  The Redis URL for Flask Cache. By default, this will REDIS_URL as set from Heroku. If you want to use a different cache, place that URL in the environment.