Repository for generic cicd templates.

Importing workflows

To import workflows, copy the examples/workflows directory to you repo .github dir on main branch.

While doing initial ci commits, include a [ci skip ] in commit message, to skip running the build-and-deploy workflow.

Trigger conditions can be set per repo in workflow file

How to set triggers

Defining k8s env

k8s_env variable in workflow defines which dir under k8s will be used as base for deployment.

e.g.

k8s_env: stage

will make cicd use k8s/stage dir

k8s_env: prod

will make cicd use k8s/prod dir

k8s_env: leftmywaller

will make cicd use k8s/letmywallet dir

Build and deploy

Runs either on push to main branch, or on manual trigger.

Steps:

- extract version from package.json / cargo.toml - file specified in cicd_inputs as version_file
- if a docker image tagged with this version hasn't been pushed to registry, build it
- if a git tag with this version doesn't exist, tag the commit 
- deploy k8s manifests, using version read from version_file
- if etcd-pull is true, pull config from etcd and inject it into k8s secret, if changed restart deployment by adding checksum annotation

If there was a change in version, but not in manifests - job will build and run a new version.

If there was no change in version, but there was a change in manifest - job will run new manifests with the latest version.

If there was no changes in manifests, nor in version, job will do nothing.

Deploy

Runs on manual trigger and requires a docker image tag to run.

Deploys given image to k8s

Dockerfile

To build a dockerimage a Dockerfile is needed.

It has to situated in examples/docker dir

and be named in convertion of

Dockerfile.app_name

corresponing to app_names provided in cicd-inputs

e.g.

Dockerfile.example

in case of monorepo, multiple Dockerfiles have to be provided, corresponding to app_names e.g.

Dockerfile.example1
Dockerfile.example2

Importing K8s templates

To enable k8s deploy, copy the examples/k8s directory to you repo root dir.

cicd-inputs file specifies variables to be passed to following pipelines and has to be set.

In case of monorepo it required to specify app_name list in format:

app_names: ["example1","example2"]

or in case of single app

app_names: ["example1"]

Manifest files have to named in convention of app_name-object.yaml e.g.

example-deployment.yaml

example-ingress.yaml

Enabling etcd pull

To enable pulling configuration from etcd, following line have to be added / uncommented from cicd-inputs:

The key path will be determined by namespace and app name e.g

/example_namespace/example_app/envs/example_key

pull_etcd_config: true              # toggle etcd pull flag
etcd_key: .env.stage                # key to be pulled 

The file to be pulled has to be in key=value format coded as yaml.

etcd key example

Defined key will be pulled from etcd during workflow execution, and saved to a k8s secret named as {{ app_name }}-secret.

To mount secret as envs, add / uncomment follwing lines from deployment's manifest:

  envFrom:
    - secretRef:
        name: example-secret