This repo contains a connection of basic Sensu assets to that can be used as a base to run a simple Sensu server.
Assets are organised by script language, so that assets can be build consistently, and in their own way where necessary
See the sensu-asset-builder repo for a way to automate the packaging of these scripts into actual Sensu assets that can be directly imported into Sensu.
The below assumes the use of the sensu-asset-builder code to actually build the assets.
Each asset needs to have it's own metadata. This is so that when building, and dependencies can be downloaded, as well as defining different operating systems (where packages may need compiling).
Below is an example asset_metadata.json file for an example asset:
{
"asset_name": "core/my_core_sensu_check", // suffix must match directory name
"description": "An example sensu check asset",
"version": "1.0",
"type": "python",
"python_version": "3.9.10",
"requirements": ["boto3"], // This should not include any modules that there's already a runtime package for
"systems": [
{
"os": "linux",
"arch": "amd64",
"platform_family": "rhel",
"platform_version": 7,
"sensu_filters": [
"entity.system.os == 'linux'",
"( entity.system.platform_family == 'rhel' || entity.system.platform_family == 'centos')",
"parseInt(entity.system.platform_version.split('.')[0]) == 7"
]
},
{
"os": "linux",
"platform_family": "alpine",
"sensu_filters": [
"entity.system.os == 'linux'",
"entity.system.platform_family == 'alpine'"
]
}
]
}The platform_family and platform_version attributes define how the operating system used to build the assets, and determine which docker image will be used.
The sensu_filters attribute uses Sensu query expressions to specify which Sensu agents each asset package will be deployed to, and hence which ones will be built. This is only important when there are compiled files within the package, and is unnecessary if the asset contains just a simple script.