Opsani CLI is cloud optimization in your terminal. It brings a suite of tools for configuring & deploying servos, managing optimization runs, and interacting with the optimization engine to your command line.
Opsani CLI is an early stage project in active development.
We need your feedback to determine how best to evolve the tool.
If you run into any bugs or think of a feature you'd like to see, please file a ticket on GitHub.
The CLI currently supports controling optimization lifecycle, managing configuration, and building servo artifacts for connecting new apps to Opsani for optimization.
To perform any meaningful work, you must first initialize the client via opsani init and supply details about your optimizer and API token.
Once initialized, you can work with the optimizer using the subcommands of opsani optimizer.
Help is available via opsani --help.
Opsani CLI respects a few key environment variables, when present:
| Variable | Description |
|---|---|
OPSANI_OPTIMIZER |
Sets the optimizer to interact with |
OPSANI_TOKEN |
Sets the API authentication token |
OPSANI_BASE_URL |
Sets the base URL for reaching Opsani API |
OPSANI_PROFILE |
Sets the profile to use (implies an optimizer, servo, token, and base URL) |
The Opsani CLI is designed to be a flexible utility that is useful in day to day development & administration activities and suitable for integration into automated workflows. The CLI can maintain a persistent configuration in ~/.opsani or can be configured via environment variables and CLI arguments for automation tasks.
Certain functionality such as the profile and servo registries are dependent on persistent storage and will emit an error if they are unavailable due to a configuration file not found condition.
To support users who work across a number of Opsani implementations, the CLI supports
named profiles. When the client is initialized, the first profile is named "default"
and is auto-selected when no profile argument is supplied. Profiles can be managed via
the opsani profile subcommands.
The primary source of documentation at this stage is this README and the CLI help text.
Opsani CLI is distributed in several forms to support easy installation.
Builds for macOS systems can be installed via Homebrew:
$ brew tap opsani/tap
$ brew install opsani-cliBuilds for Windows systems are available via scoop:
$ scoop bucket add opsani https://github.com/opsani/scoop-bucket.git
$ scoop install opsani-cliNOTE: Opsani Ignite requires a hypervisor to run minikube virtualization. Microsoft provides the Hyper-V hypervisor for modern version of Windows that can be installed via PowerShell. Install Hyper-V using the instructions at the Microsoft Hyper-V Quick Start page.
Binary packages for Linux are available in RPM and DEB package formats. Download the appropriate package for your distribution of the latest release and then install as detailed below.
$ sudo apt install ./opsani-cli_*_linux_amd64.deb
$ sudo dnf install opsani-cli_*_linux_amd64.rpm
$ sudo yum localinstall opsani-cli_*_linux_amd64.rpm
$ sudo zypper in opsani-cli_*_linux_amd64.rpm
Versioned binaries for all platforms are distributed via GitHub: https://github.com/opsani/cli/releases.
Download and install the appropriate build for your platform
The Makefile is configured with a make install target that will build and
install the CLI into /usr/local/bin/opsani on Unixy platforms.
Containerized releases of Opsani CLI are pushed to the Opsani organization on Docker Hub. To retrieve the latest container image:
$ docker pull opsani/cli:latest
or
$ make imageOnce the image is pulled, configuration must be provided to enable the CLI to function. There are a couple of options for supplying the configuration:
- CLI Arguments - Arguments can be directly supplied to the CLI via the Docker invocation:
docker run -it --rm --name opsani-cli opsani/cli:latest --optimizer example.com/app --token 123456 - Environment Variables - Export
OPSANI_OPTIMIZERandOPSANI_TOKENinto your shell and pass them through to the container:docker run -it --rm --name opsani-cli -e OPSANI_OPTIMIZER=$OPSANI_OPTIMIZER -e OPSANI_TOKEN=$OPSANI_TOKEN opsani/cli:latest - Config Volume - Existing configuration files in your home directory can be shared with the container through a
volume mount:
docker run -it --rm --name opsani-cli -v ~/.opsani:/root/.opsani opsani/cli:latest
Commands can then be appended to the end of the command just as when executing in a local shell.
Note that running under a container for day to day work is typically unnecessary and inconvenient as Opsani CLI is distributed as a single statically linked binary.
This repository includes a Dockerfile for building and running Opsani CLI in a container. To build the image from the Dockerfile:
$ docker build . -t opsani/cli:latestOpsani CLI is implemented in Golang and utilizes a handful of well established libraries including Cobra, Viper, and Resty. Anyone reasonably well versed in Go should have no trouble following along.
There is a Makefile for running typical tasks but go run . is a great way to
poke around.
Opsani CLI has extensive automated test coverage. Unit tests exist alongside the
code under test and there are integration tests in the integration directory.
Testing assertions and mocks are provided by the
Testify library. Tests are organized
using the testify.suite package to provide familiar xUnit testing primitives.
The local test package contains shared testing helpers useful in both unit and integration test scenarios.
Tests can be run via the Makefile:
make test- Run unit & integration tests.make test_unit- Run unit tests.make test_integration- Run integration tests.
The integration test harness functions by building the opsani binary, copying
it into a temp directory, and then interacting with it as a subprocess and
evaluating stdout, stderr, and exit code status.
API interactions are mocked locally via an http.httptest server that runs in
the test executor parent process. The --base-url flag is used to direct the
CLI to interact with the test API server rather than the Opsani API.
Opsani CLI is released under the Apache 2.0 license. See LICENSE