A generic service health sidecar
Find a file
2024-07-25 03:35:01 +00:00
.devcontainer chore: upgrade to Go 1.22 2024-04-11 17:23:47 +02:00
.forgejo/workflows chore(deps): update goreleaser/goreleaser-action action to v6 2024-06-09 03:34:15 +00:00
api feat: refactor to server and exec-check subcommands 2023-12-04 11:22:49 +01:00
check chore: cleanup, add docs, refactor some quirks and prepare release 2023-12-04 16:59:10 +01:00
cmd fix: golangci-lint settings 2024-04-10 08:35:15 +02:00
config fix: golangci-lint settings 2024-04-10 08:35:15 +02:00
deployments Add basic Redis checks 2022-04-28 18:35:02 +02:00
grammar refactor: migrate to new Gitea URL 2022-09-22 11:46:36 +02:00
internal chore: cleanup, add docs, refactor some quirks and prepare release 2023-12-05 07:19:39 +01:00
magefiles feat: refactor to server and exec-check subcommands 2023-12-04 11:22:49 +01:00
protocols chore: cleanup, add docs, refactor some quirks and prepare release 2023-12-04 16:59:10 +01:00
validation feat: add concourse pipeline 2022-12-30 21:41:30 +01:00
.editorconfig feat(ci): move to Drone CI 2023-02-22 22:26:29 +01:00
.gitignore feat(ci): display test results in drone 2023-03-05 11:28:20 +01:00
.golangci.yml fix: golangci-lint settings 2024-04-10 08:35:15 +02:00
.goreleaser.yaml feat(ci): use ko for image builds 2023-12-12 21:21:12 +01:00
.husky.toml feat: refactor to server and exec-check subcommands 2023-12-04 11:22:49 +01:00
go.mod fix(deps): update module github.com/urfave/cli/v2 to v2.27.3 2024-07-25 03:35:01 +00:00
go.sum fix(deps): update module github.com/urfave/cli/v2 to v2.27.3 2024-07-25 03:35:01 +00:00
LICENSE Initial commit 2022-04-13 07:39:19 +02:00
main.go feat: refactor to server and exec-check subcommands 2023-12-04 11:22:49 +01:00
nurse.yaml Implement config reading 2022-05-08 11:00:22 +02:00
README.md chore: cleanup, add docs, refactor some quirks and prepare release 2023-12-04 16:59:10 +01:00
renovate.json Add renovate.json 2022-04-13 05:39:26 +00:00

Nurse

Usage

Nurse comes currently with 2 different operation modes:

  • server
  • CLI

The server starts an HTTP server with configurable endpoints which you can use e.g. in Kubernetes environments to distinguish between:

  • startup
  • readiness
  • liveness

probes. Every endpoint has a distinguished set of checks that are executed when you hit the endpoint. Currently, there is no caching in place (and there are also no plans to change that).

The CLI operation mode on the other hand executes all checks that are provided as arguments e.g. in Docker Swarm environment where the container image has to ship the health check CLI.

Primer about checks

All checks are executed in parallel which means you shouldn't rely on a certain execution

Global config/options

Nurse comes with the following global options:

Switch Environment variable Default value Description
--config NURSE_CONFIG $HOME/.nurse.yaml, /etc/nurse/config.yaml,./nurse.yaml path to the config file
--check-timeout NURSE_CHECK_TIMEOUT 500ms Timeout for executing all checks
--check-attempts NURSE_CHECK_ATTEMPTS 20 How often checks should be retried before they're considered failed
--log.level info Default log level
--servers NURSE_SERVER_<name> Configure server URLs via environment variables

The individual sub-commands come with additional options, like for example configuring endpoints via environment variables as well.

The nurse.yaml describes how to configure Nurse via a configuration file.

The most interesting root nodes are:

  • servers
  • endpoints

Within servers you can configure different servers for further usage in checks. For example, to configure a Redis server: redis://localhost:6379/0. Depending on the individual protocols there are further configuration options.

Within endpoints you can configure different HTTP endpoints the server exposes and which checks should be executed for which endpoint.

Server

The server sub-command comes with the following additional config options:

Switch Environment variable Default value Description
--endpoints NURSE_ENDPOINT_<name> Configure HTTP endpoints via environment variables
--http.address NURSE_HTTP_ADDRESS :8080 IP and port the server will be listening on
--http.read-header-timout NURSE_HTTP_READ_HEADER_TIMEOUT 100ms Timeout until when the client has to have sent the headers

To configure an endpoint via an environment variable, set it like this:

NURSE_ENDPOINT_HEALTHZ='http.GET("https://api.chucknorris.io/jokes/random")=>Status(200);redis.PING("local-redis")'

The server will print the configured routes when it is starting up. In the aforementioned case you should see something like:

{"time":"xxxxx","level":"INFO","msg":"Configuring route","route":"/healthz"}

Multiple checks can be configured by separating them with a ; into multiple 'expressions'.

CLI

The CLI has no additional config options compared to the server. It simply takes all arguments you pass to it, tries to parse them as checks and executes them with the given time limit. If one of the check fails it will exit with a non-zero exit code.

Multiple checks can either be passed as single argument in '' separated with a ; just like in the environment variables, or you can pass multiple arguments. The result will be the same.