CI Plan

Background

The sBTC bridge and API applications have been developed so far on two separate stacks for the staging and production environments.

We currently use Linode (Akamai) for the staging build/deployment process and Google Cloud Platform for the production and CI environments.

This supported the initial prototyping and bootstrapping phase of the Bridge project but maintaining this going forward will dilute the skillsets and focus required to set up a robust CI for the Bridge project.

Solution architecture requirements;

• Geographical load balancing of client and api applications
• Support for multiple networks (testnet, mainnet)
• Support for full, multiple blockchain nodes (stacks, bitcoin)
• Redundancy in the data layer
• Scalability of the API and DB layers
• Backup / restore data features

CI Definition

Branches for building the staging and production branches;

• staging for staging
• main for production

Client CI

1. Client deploys on PR merge to staging branch assuming all CI checks pass
2. Staging client has stable ssl URL e.g. https://staging.sbtc.world
3. Client deploys on PR merge to main branch assuming all CI checks pass
4. Production client has stable ssl URL e.g. https://sbtc.world
5. Staging and production clients are hosted in GCP
6. Transient development build (PRs etc) use Cloudflare for running integrity checks

API CI

1. The API node express app runs in GCP Kubernetes cluster
2. The cluster provides scalability, redundancy and availability of the api as needed via configuration.
3. Scalability, redundancy and availability of the db are provided using Mongo Atlas service provider.
4. GCP supports integration with Atlas

Notes

Bridge Web

• Static files
• Can use any service
• Uses GCP Buckets for Production
• Uses CloudFlare for Developer branches
• Load Balancer -> reads Bucket

Deployment API

• Dockerfile.api
• Cloud Run or K8s
• CHANGES(mijoco)
• Fork Dockerfile into Dockerfile.api
• Read chain from env variable
• Disable time here

Deployment Functions

• Dockerfile.lambda
• Google Lambda or K8s
• CHANGES(mijoco)
• Fork Dockerfile into Dockerfile.timer
• Read chain from env variable
• Disable the http request handler