Kentik platform API schemas and pre-built code

About Kentik APIs

The Kentik platform currently provides 2 generations of public APIs:

• legacy REST-only APIs, known as API version 5
• modern gRPC and REST APIs, known as API version 6

This repository contains schemas and pre-built artifacts only for modern, version 6 APIs.

More information about Kentik APIs (including legacy version 5 APIs) is available in Kentik Knowledge Base.

Version 6 APIs

Modern Kentik APIs are based on gRPC utilizing Protocol Buffers for transfer encoding. All API schemas are defined in proto files using version 3 syntax.

REST/JSON APIs are also provided via automatic API translation using gRPC-Gateway. OpenAPI (a.k.a. Swagger) version 2 and 3 spec files are provided for all version 6 APIs.

API versioning

Kentik APIs use Calendar Versioning with following format:

v<YYYY><MM>[(alpha|beta)][<N>]

where:

• YYYY is year number (e.g. 2022)
• MM is month number (e.g. 02, 11)
• alpha or beta suffix indicating life-cycle status of the API (see bellow)
• N optional serial number (used if more than 1 version must be introduced in the same month)

API life-cycle status

APIs are in 1 of 3 possible stages based on their life-cycle status:

• alpha: experimental, incomplete and mostly untested APIs
• beta: functional, usable, but potentially incomplete APIs
• stable: mature API with the highest level of support

Suggestions for use of unstable APIs:

• alpha APIs shall not be used for any production purposes
• beta APIs shall not be used for any critical use cases requiring long term support.

The table bellow provides service level objectives for APIs based on their life-cycle status

life-cycle status	backward compatibility within a version	support for old versions
stable	guaranteed	12 months after release of new version
beta	guaranteed	6 months after release of new version
alpha	not guaranteed	none

Stable APIs become deprecated as soon as a new version is released. Deprecated versions do not get any futher development beyond critical bug fixes. Most importantly, they do not receive support for new features.

New features, extensions of existing functionality may be introduced in stable APIs as long as added functionality is backwards compatible (i.e. does not allow any changes in behavior of clients).

Repository layout

• proto - Kentik API schemas
• vendor proto - API schemas utilized by Kentik APIs
• gen - pre-built artifacts
  • proto descriptor file containing all APIs and dependencies
  • pre-built API client and server stubs in
    • Go
    • Python
    • C
    • C++
  • OpenAPI specs/gen
    • v2
    • v3

Using the generated code

Go

The repository is configured to provide installable Go package named github.com/kentik/api-schema-public. Individual APIs

import <alias> "github.com/kentik/api-schema-public/gen/go/kentik/<api_name>/<api_version>"

It is recommended to always explicitly specify the import alias. If no alias is specified, API objects are imported with name specified in the last component of the go_package option value (after ;) in the corresponding proto file. For example the go-package option of the Synthetics API is github.com/kentik/api-schema-public/gen/go/kentik/synthetics/v202202;synthetics, so the following code is valid:

import (
	"fmt"
	"github.com/kentik/api-schema-public/gen/go/kentik/synthetics/v202202"
)

func main() {
	fmt.Println(synthetics.TestStatus_TEST_STATUS_ACTIVE)
}

Python

The repository is configured to allow installation of package named kentik-api-proto containing code for all Kentik APIs directly from GitHub using pip:

pip install git+https://github.com/kentik/api-schema-public.git

The repo-spec string can also be used in requirements.txt. Locally the package can be built in the checked out repo with:

python -m build

which creates Python "wheel" package in the dist/ directory.

With the kentik-api-proto package installed, protobuf messages and gRPC client stubs can be imported from the kentik namespace. Examples:

from kentik.synthetics.v202202.synthetics_pb2 import ListTestsRequest
from kentik.synthetics.v202202.synthetics_grpc import SyntheticsAdminServiceStub

The Python code requires the grpclib package which is installed as dependency.

Caveats:

1. As of now, the Python code requires PROTOCOL_BUFFERS_PYTHON_IMPLEMENTATION=python to be set in the environment. This will be fixed in the near future.
2. Using any gRPC services requires asyncio style programming. Aall methods of <service>Stub classes return asyncio futures.