ibm-cloud/insurance-toolchain

Toolchain configuration files for the insurance-store microservices demo

insurance-toolchain's Issues

Make name of deployed apps unique

When creating the toolchain, "standard names" for the 4 microservices are chose.

There is no warning to users if these routes are already taken. Errors can be found in the logs:

The route dev-cloudco-insurance-catalog.mybluemix.net is already in use.

But these are hard to find or remedy (for novice cloud foundry users)

Is it possible to ask the user for a pre- or suffix for the apps? Or (at the very least) add a prominent note to the README what needs to be done to avoid this situation?

SOCKET_URL is not set at the right time in the pipeline

SOCKET_URL is configured during DEPLOY but it is really the BUILD phase that needs it.

Update live toolchains, split in two dev and master

Following #12, update the live toolchains with two chains.

Allow to deploy only master or only dev

The current toolchain deploys both dev and prod versions of the app. What if I'm only interested in the master stream?

What if I want to deploy master for all modules but dev for the dashboard as I want to do fixes there?

Instead of deploying master+dev in one toolchain, we could have the option to select which branch to deploy for each module, defaulting to master.

Remove extra bind service in the DEPLOY stage of (prod) pipelines in the toolchain

They need to exist until the changes made to make the README.md usable are merged from dev to master. They should work even after merge but will be redundant with the auto-bind made by the manifest.yml

Add bot URL to the dashboard as an env variable

Dashboard needs the URL of the bot:

socket_url : process.env.SOCKET_URL || 'dev-cloudco.mybluemix.net',

This task is for the toolchain to inject this env var

Authentication failure in the DEPLOY stage

Hi. I'm seeing an error with both pipelines (bot and dashboard). The BUILD stage was successful, but the erorr below was found on the DEPLOY stage:

---

Preparing to start the job...
Preparing the build artifacts...
cf login -a "https://api.eu-gb.bluemix.net" -u "" -p "" -o "" -s ""
API endpoint: https://api.eu-gb.bluemix.net

Email>
Password> stty: 'standard input': Inappropriate ioctl for device

stty: 'standard input': Inappropriate ioctl for device
Authenticating...
Credentials were rejected, please try again.

Password> stty: 'standard input': Inappropriate ioctl for device

stty: 'standard input': Inappropriate ioctl for device
Authenticating...
Credentials were rejected, please try again.

Password> stty: 'standard input': Inappropriate ioctl for device

stty: 'standard input': Inappropriate ioctl for device
Authenticating...
Credentials were rejected, please try again.

API endpoint: https://api.eu-gb.bluemix.net (API version: 2.106.0)
Not logged in. Use 'cf login' to log in.
FAILED
Unable to authenticate.

Finished: FAILED

---

Adapt toolchain to demo script

Change toolchain so that:

• All 9 microservices deploy into a single space (the one that the user is logged into). Services will have different name prefixes. Example:
• dev-{service-name}
• test-{service-name}
• {service-name}
• Remove slack integration (will be added interactively during the demo)

Record screencast of demo

Record screencast and add link to README.md

Deploy failed

I have this error message in the log
FAILED
The plan Lite could not be found for service cloudantNoSQLDB

Free Trial comes with only 2 GB runtime memory

Toolchain seeks 3.2 GB of free memory for deployment. But Bluemix Trial account provides only 2 GB of Runtime + Container memory . Check pricing

Move all pipeline scripts to their respective repo

Instead of having the scripts inside the pipeline yml files, move them in their respective repos. It makes updating the deployment logic easier with less updates to do in the pipelines.

See logistics-wizard-recommendation for an example.