Download the PHP package oniva/flow-healthstatus without Composer
On this page you can find all versions of the php package oniva/flow-healthstatus. It is possible to download/install these versions without Composer. Possible dependencies are resolved automatically.
Download oniva/flow-healthstatus
More information about oniva/flow-healthstatus
Files in oniva/flow-healthstatus
Package flow-healthstatus
Short Description Flow Framework package to check the health status of you application
License MIT
Informations about the package flow-healthstatus
Oniva.Flow.HealthStatus
Package to check the health status of a flow application.
It's extremly useful in a kubernetes environment to use with readiness and liveness probe to determine if a pod can serve traffic and if it is still alive.
Usage
To determine the current health status of your application you can check wether the app is ready or still alive.
Readiness
Simply execute the flow command ./flow app:isready.
This will execute all tests defined in the Oniva.Flow.HealthStatus.testChain of the Settings.yaml.
If all tests have passed, the readyChain tasks will be executed.
After a successfully run of the readyChain an internal lock will be set to prevent repeated execution. The testChain will be executed on every run.
So the readyChain should bring your application in an "ready state". Make sure to initialize everything you need.
The testChain should ping all services your application depends on.
Liveness
Execute ./flow app:isalive to check if your pod is still alive.
This will execute the livenessChain.
Currently the liveness chain is empty by default and has one possible test: statusCode.
Configuration
Add all your tests in the following format in your apps Settings.yaml:
After that, the check will execute the ready chain:
After a successful ready chain invokation, you can call ./flow app:isalive to execute your liveness chain:
Advanced configuration
Before each attempt to execute a ready task,
the check will test the Oniva.Flow.HealthStatus.defaultReadyTaskCondition to see if the task should be executed.
In the default configuration this is simply a check to see if the ready lock is not yet set.
You can override this behaviour on a per task basis:
(the lockName setting is simply a shorthand for exactly this example)
To extend the eel context, you can provide additional helpers in Oniva.Flow.HealthStatus.defaultContext.
Health status via HTTP-Request
If you'd like to check your Application via HTTP instead of a cli command you can do so by including the Routes in your Routes.yaml:
Afterwards there will be two new routes: /<your-endpoint-name>/ready and /<your-endpoint-name>/live
Adjust the variable to your needs. Both endpoints will return a JSON formatted output. On a successful run the
response has a status code 200 and if there are any errors the status code will be 500.
Example Configuration
This example could be used in your Flow package to make sure that your application pod has a ready state to serve traffic. Therefore it will always check the ping status for doctrine, redis and beanstalk. On the first run all missing database migrations will be executed, the redis cache flushed and static resources published. After a successfull run only the testChain will be executed again.
Note the lockname configuration. This Configuration enables you to run tasks only once per deployment or always.
By default the Oniva_FlowHealthStatus_Lock cache is used to read and write locks. Add this to your Caches.yaml and all your application pods will
rely on the same lock files as they don't use the local file storage but redis. This will result in a execution once per deployment:
The staticResources task has a custom cacheName configured. To ensure that this task will be executed in each application pod set it to local file storage:
Migration from t3n.Flow.HealthStatus
There is a code migration script to migrate your existing t3n.Flow.HealthStatus installation to Oniva.Flow.HealthStatus.
Simply run the flow command:
Acknowledgments
This project is a fork of t3n.Flow.HealthStatus originally developed by t3n. We thank the original contributors for their foundational work.