build-env

This is a docker image for ci building

It is based on a python3 image and the following programs

deploy-versioner

A program for automatic setting new versions in kubernetes deployment configurations, meant for a gitops workflow.

https://github.com/DBCDK/kube-deployment-auto-committer

Help

$ set-new-version -h
usage: set-new-version [-h] [-b BRANCH] [--gitlab-url GITLAB_URL] [-n]
                       deployment-configuration gitlab-api-token project-name
                       image-tag

positional arguments:
  deployment-configuration
                        filename or dir of deployment configuration to change.
                        If a dir is specified, all .yaml and .yml files in
                        that dir recursively will be included.
  gitlab-api-token      private token for accessing the gitlab api
  project-name          Name of project (including group hierarchy), e.g.,
                        metascrum/rrflow-deploy
  image-tag             new tag to commit into the deployment configuration

optional arguments:
  -h, --help            show this help message and exit
  -b BRANCH, --branch BRANCH
  --gitlab-url GITLAB_URL
  -n, --dry-run         don't commit changes, print them to stdout

webservice-validation

A program for validating responses from a webservice.

The validation spec must be yaml and consist of a list of validation objects. The format will likely change slightly as we begin to use it and discover shortcomings.

The specific part of the response to be validated (at the moment validation means checking if the object has a length equal to or greater than the value specified by response.len in the validation spec with a default of 1) can be extracted with jsonpath using response.jsonpath. All of the jsonpath specification is supported, but only the most simple use-cases are tested as of the initial commit.

• https://goessner.net/articles/JsonPath/
• https://pypi.org/project/jsonpath-ng/

To test against multiple endpoints you can supply a json file with a list of endpoints to test. The format must be a list of objects containing an ip key with the endpoint. This output can be fetched from the kube-tools ip-addresses command (https://gitlab.dbc.dk/ai/kube-tools-rs). It looks like this

[
    {
        "ip": "10.233.22.82",
        "name": "subject-suggester-1-0-7c785b866c-gp5dz"
    },
    {
        "ip": "10.233.33.145",
        "name": "subject-suggester-1-0-7c785b866c-ltxc7"
    },
    {
        "ip": "10.233.2.143",
        "name": "subject-suggester-1-0-7c785b866c-ss82s"
    }
]

When testing with output from the kube-tools command be aware if you need to also specify the port to use on the endpoint. You can do that with -p or --port.

Ex:

- validation:
  path: /api/subjects/related/emne?record_id=validation-test
  method: post
  data: '["666*fhistorie", "666*fpolitik"]'
  response:
    jsonpath: "response"
    status_code: 200
  headers: {"content-type": "application/json"}

Help

$ webservice-validation -h
usage: webservice-validation [-h] [--from-endpoints-json-file FROM_ENDPOINTS_JSON_FILE] [-p PORT] endpoint validation-spec

positional arguments:
  endpoint
  validation-spec

optional arguments:
  -h, --help            show this help message and exit
  --from-endpoints-json-file FROM_ENDPOINTS_JSON_FILE
                        Validate from a list of endpoints instead of the single positional argument endpoint. The format should be a list of object containing an `ip`
                        key with the endpoint. This output can be fetched with the `kube-tools ip-addresses` command.
  -p PORT, --port PORT  Port to query on the service which is to be validated.

Examples

kube-tools ip-addresses mi-staging app=simple-search-1-0 | webservice-validation from-endpoints-json-file deploy/validation.yml --from-endpoints-json-file - -p 5000