It would be nice to have access to the RSA Public key in the project settings page

Following up after #26

How about supporting a config file (.codeflow.yml) to define resources, and other project definition (extensions too?).

We should prevent users from saving 0 as port number

When changing type from internal to external delete the load balancer and then create a new one or lock the option and let user delete and create

needed to run migrations

When multiple deploys are pending, the order of actual release is dependent on image build and CI times.

If multiple deploys for the same projects are pending, we should cancel the oldest ones and deploy only the last one.

For example, deploy at 5PM everyday.

How to reproduce

• Get into a project that has at least 2 resources
• Click edit on one of the resource (Do not change save or click anything)
• Click edit on a different resource

Expected:
The input values are changed to the resource being edited

Actual
Inputs become blank

go test `go list ./... | grep -v vendor`

have lots of failures.

We may want to fix them and integrate the go test into CI.

Feature request: If an admin github user was authenticated, it should be able to add deploy keys, webhooks, etc for a project automatically.

As per suggestion from nampnq on slack:

I think should ignore migrate route53 if not config, else it will raise error when make connect to route53

https://github.com/checkr/codeflow/blob/master/server/plugins/codeflow/migrations/mongodb.go#L927

This would allow us to deploy any project from git 🚀

We can use heartbeat plugin to git pull every minute and create new events for Codeflow.

1. Codeflow receives heartbeat
2. Codeflow emits pull events
3. Plugin pulls code and emits events for new features
4. Codeflow stores state, builds, deploys...

Feature request:

Incremental rollout and host level deployment. The finer granularity of the deployment flow enables more flexible rollout, for example, testing in production with pinned hosts, datacenter specific deployment, and incremental rollout with alert triggered rollback.

Immediate need:

• set/unset env var
• deploy

Nice to have:

• rollback
• scale service

I'm not sure how exactly this should be handled, but, if you can have duplicates, we should allow for configurable project names. Otherwise, we should restrict a project being added at most once (which kind of sucks if you want to run in different environments).

This will provide better feedback!

Lodash seems to be the preferred build now that underscore and lodash communities have been merged. Lodash also supports importing individual functions which can cut down on build sizes.

e.g.
import { each } from 'lodash'

We should timeout on CI builds, and in case of timeout fail the deploy on codeflow (which would notify slack!).

When modifying environment variables, enter should trigger the save button and not the delete project button. The dialogue also gives no context that you are about to delete a project so it's extremely easy to accidentally delete it

Currently when doing POST, PUT, DELETE we just redirect to GET endpoint and return full response. This was a shortcut to get things moving fast but it's incorrect and should be fixed

Ex: https://github.com/checkr/codeflow/blob/master/server/plugins/codeflow/bookmarks.go#L77

Run multiple commands for a service as separate containers on the same node. This would allow us to use PODs to share volumes between 2 or more services.

In the interface/service there could be a + next to a command that adds another command field.
We should also support adding volume paths that containers are sharing...

Instead of rolling back it should just send the failure message to codeflow and let it do the rollback.

In Codeflow dashboard switch to another environment that keeps track of its own features, services, releases and has its own settings. Used for staging and development environment and with combination of issue #27 we could use it to deploy and test PR's

The setting is not changed even after hitting save button.