Health Check
This documentation will introduce how to use health
policy to apply periodical
health checking to an application.
#
BackgroundAfter an application is deployed, users usually want to monitor or observe the health condition of the running application as well as each components. Health policy decouples health checking procedure from application workflow execution.
It allows to set independent health inspection cycle, such as check every 30s. That helps users to notice as soon as applications turn out unhealthy and follow the diagnosis message to troubleshot.
#
Health PolicyWe apply the sample application including two components, my-server
is
supposed to be healthy while my-server-unhealthy
is supposed to be unhealthy
(because of invalid image).
As shown in the sample, a Health
policy is specified.
Health
policy accepts two optional properties, probeInterval
indicating time
duration between checking (default is 30s) and probeTimeout
indicating time
duration before checking timeout (default is 10s).
To learn about defining health checking rules, please refer to Status Write Back.
Finally we can observe application health status from its status.services
.
Here is a snippet of health status.
#
Appendix: Parameter ListName | Desc | Type | Required | Default Value |
---|---|---|---|---|
probeInterval | time duration between checking (in units of seconds) | int | false | 30 |
probeTimeout | time duration before checking timeout (in units of seconds) | int | false | 10 |