- Get health statuses of all services: http://localhost:8080/api/healthStatus/getHealthStatuses
- Response is an array of service statuses.
- Each status will have a timestamp, code and isOk.
- Get services availability: http://localhost:8080/api/healthStatus/getServicesAvailability
- Response is an array of service availability percentages.
The settings.js file contains a list of services to monitor. Each service in the list is represented by an object that contains the service name, uri, responseType (xml / json), statusPath (the object path to the property that represents the overall status) and okWord (the term that symbols if status is OK).
This file contains environment variables (server port, interval value, buffer max size).
My main focus was on how to store the logs for each service and how to parse thier responses. I wanted to make it scalable so it can be expanded to more then 3 services.
The main problem was that each service response in a different structure and that's why I had to describe each service in the settings.js file.
Each service has its own HealthBuffer. The HealthBuffers are stored in a global singleton object.
I tried to split the code to small components to ease on reuse, make the code readable and to ease on unit testings.
-
Logs are pushed to a FIFO on each interval and therefore the push might be inaccurate - if the interval is delayed, an early log can get pushed to the fifo after a later one.
-
On first minute - the buffers will be empty cause it takes the main code 1 minute to run because of the interval.
Run npm test
Run npm install and then npm start