Probe

Probe is a health check service that can do health checking for multi DB engines, running as a sidecar in cluster pods.

1. Introduction

Probe is capable of doing multi checks, include running check/status check/role changed check, serviced through HTTP API. we use kubelet readinessprobe to config and enable probe in kubeblocks.

2. Getting Started

You can get started with Probe, by any of the following methods:

• Download the release for your platform from github.com/apecloud/kubeblocks/release
• Use the available Probe docker image docker run -it apecloud/kubeblocks
• Build probe from sources

2.1 Build

Compiler Go 1.20+ (Generics Programming Support), checking the Go Installation to see how to install Go on your platform.

Use go build to build and produce the probe binary file. The executable is produced under current directory.

$ cd kubeblocks/cmd/probe
$ go build -o probe main.go

2.2 Configure

Probe read some configurations from env variables, Now there are: KB_CHECK_FAILED_THRESHOLD: The number of continuously checks failed aggregated in one event, in case of too many events sent. minimum 10, maximum 60 and default 10. KB_ROLE_UNCHANGED_THRESHOLD: the count of consecutive role unchanged checks emit one event. minimum 10, maximum 60 and default 60. KB_SERVICE_PORT: The port of service to probe, eg. 3306. KB_SERVICE_USER: The user name of service used to connect, eg. root. KB_SERVICE_PASSWORD: The user password of service used to connect. KB_SERVICE_ROLES: The mapping of roles and accessmode, eg. {"follower":"Readonly","leader":"ReadWrite"}.

2.3 Run

You can run the following command to start Probe once built

$ probe --config-path config/probe --port 7373

3. License

Probe is under the AGPL 3.0 license. See the LICENSE file for details.