My Reminders for Cloud

Important: This repository is in the process of moving to Github. Until the move is finalised, master in Bitbucket will be synced with master in Github at a regular cadence.

The My Reminders plugin for JIRA (and eventually the other products too) is an Atlassian Connect addon that has been designed to complete the notification flow that is missing in Atlassian products. You can be notified of changes to things that you are interested in but you don't get notifications for items that do not change.

Every now and then you want to be notified about an important issue that MUST be completed before a particular date, for example, if another team promised you that a particular issue would be resolved by a certain date then you would really like to be reminded to check up on that issue by the given date. That is what the My Reminders plugin for JIRA solves, it will remind you of an issue in a given amount of time.

Screenshots

Create a Reminder

Recieve an Email Reminder

Manage your Reminders

Developing the Atlassian Connect Addon

Database setup

You'll need a postgreSQL server available, matching the access configurations in snaplets/postgresql-simple/devel.cfg.

There is a convenience script at schema/bootstrap.sh to set up the database for you, assuming that the current user has permission to create databases.

During development, if you make changes to migrations that have already run locally, you'll need to the dbmigrations moo tool. You can install it by running:

stack install dbmigrations-postgresql

And then you can use the tool by running:

yarn moo --help

Please don't modify previous migrations, only add new ones.

Running the plugin

First setup a nix-shell environment with the haskell stack command:

nix-shell -p postgresql zlib jdk stack libiconv

Then use the stack command to build the service:

stack build

To continuously build the service while developing:

stack build --file-watch

To run the tests:

stack test

And to run the service:

stack run my-reminders

To run the service in a development watch mode:

CONNECT_BASE_URL=https://95e354fe3704.ngrok.io yarn dev-watch

You can test that the App is running by hitting: http://localhost:8000/atlassian-connect.json

Deploying with Docker

This project has opted to use Docker as the deployment mechanism. To build the development docker image run the following in the root directory of the project:

time docker build .

Using the time command means that you can get a feel for how long the docker build will take on your machine. Once you have the docker build you can make the production docker image by:

CONTAINER_ID=XXXX bash to-production.bash

Which will copy the build image from the build container and pass it through to the production container. It will also pull through all of the required resources. This production image should be possible to deploy independently of everything else. The reason that we have this separation is that you require > 2GB of image size to create the build docker container and only ~270MB of image size to create the production container. This allows us to have much more efficient production deployments.

Local testing with Docker

The addon will reject installation requests from hosts that aren't whitelisted. If your JIRA installation has a base URL with a hostname other than "localhost", the hostname of the docker image, or one of the Atlassian OnDemand public domains, you will need to modify the whitelist (found in src/Connect/Connect.hs).

You will also need to modify snaplets/postgresql-simple/devel.cfg to point to your database from the Docker container, as the database will not be running in the addon's Docker image.

Dependencies

My Reminders is built against the Haskell Platform, currently version 2013.2.0.0. The latest version of Cabal is always recommended, we have tested with at least cabal-install version 1.20.0.2.

Additional dependencies required on PATH include:

• pg_config (e.g. apt-get install libpq-dev on Debian/Ubuntu)

Running the Code

There is a useful trick where you can hit /admin/reload on localhost and refresh the entire service without having to restart the process. This is excellent for rapid development.

Optional dependencies

These are not strictly necessary but make life easier.

UPX

Compresses the executable. make dist currently strips debug symbols and compresses the executable.

» brew install upx

cabal-constraints

The make freeze target "freezes" the exact versions of the dependencies selected by cabal-install. The dependency list will be written to cabal.config. This requires the cabal-constraints executable to be available:

» cabal install cabal-constraints

Deploying this service to production

Now that you have built the docker container you are going to want to deploy this service to production. Just place the docker container somewhere and give it the following environment variables:

• http_proxy: Required if you need to route outgoing requests through a HTTP proxy.
• https_proxy: Required if you need to route outgoing requests through a HTTPS proxy.
• PG_MY_REMINDERS_HOST: The PostgreSQL host.
• PG_MY_REMINDERS_PORT: The PostgreSQL port.
• PG_MY_REMINDERS_SCHEMA: The PostgreSQL database name.
• PG_MY_REMINDERS_ROLE: The PostgreSQL role that has access to that database.
• PG_MY_REMINDERS_PASSWORD: The password of the role.
• CONNECT_BASE_URL: The base url that should appear in your Atlassian Connect Descriptor.
• CONNECT_SECRET_KEY: The secret key that Atlassian Connect will use for page tokens.
• EXPIRE_KEY: The key to pass to the service to give access to the expire handler.
• PURGE_KEY: The key to pass to the service to give access to the purge handler.
• MIGRATION_KEY: The key to pass to the service to give access to the migration handler.

If you have properly set these environment variables then the service will be up and running in more time. For greater control you may want to edit the configuration properties in the resources directory.