Spiceworks - Health Checker Assignment
- A web service that receives thousands of requests per second
- You own a web application that receives thousands of requests per second
- Each web request is evaluated concurrently
- The application depends on several external resources
- Each of those resources might be in a healthy or unhealthy state
- Each resource will be monitored (polled) at different intervals in independent threads
- Assume the health of each resource will be random and change several times throughout the life of the application
- The health of each resource will be reported to a central
HealthAggregatorusing the methodSetResource—described below—and will contain the most recent health status for a single resource - On each request to the main app, one of the steps to perform is call
HealthAggregator.IsHealthy()to determine the overall health of the application - If any one of the resources is unhealthy, your web application will respond to the web request with a 5XX code
- Only if all resources are healthy will the response code be a 2XX
- Language: Python
- Web Service: Flask
- ORM: Flask SQLAlchemy
- Database: SQLite
- Encryption: AES with RSA keys
- Encoding: SHA512
- Communications:
- Windows: PowerShell Remoting PowerShell version should be 4.0 and above
- UNIX: SSH using Paramiko
Upon launch of the WebService, we first check if the database exists or not. If the database does not exists, we create, and add the User and Resource tables. We then check if the RSA private and public keys files exists or not.
For each resource, we create an object of the Resource class, which executes the commands on the remote machine to get the details of the machine, and to check if the resource is healthy or not.
The user table is initialized with one User in the beginning, which can perform all operations.
Username: SpiceWorks
Password: HealthChecker
We are using SQLite database for this prototype.
The database has the following tables:
-
User Table to store information related to the users which are allowed to perform advanced operations
-
Resource Table to store information related to the resources which needs to be monitored
User table stores the username, password, and a special value salt.
The password is one way encrypted (encoded) for security purposes.
For authentication, we use the value of salt for the given username, and the password that the user gave, and check if the final value matches the value stored in the table.
The Resource table stores the hostname, username, password, and interval value of the resource.
Interval value is set default to 60 (in seconds).
Resource password is encrypted using AES cipher and RSA keys.
All the endpoints and their info can be retrieved by calling the /Routes or /routes endpoint.
To launch the webservice, launch command prompt as Administrator from the root directory and run:
>>> python src
OR
>>> python src\app.py
- Execute the steps under the Usage section above
- Open Postman
- Call /Login with the credentials shared under Design section
- Copy the token and pass it under headers for subsequent API calls
- Call /Resource to add a new Resource
- /IsHealthy to check the service health
- Returns HTTP Status 200, if all the resources are healthy
- Returns HTTP Status 503, if any of the resource is unhealthy
Other endpoints and their details can be retrieved using the info from Endpoints section above.
Due to limited time, I could only test it out manually, but could not write any Unit Tests.
- Get and store the reason for down state of each resource using the Resource class, and store the info in Health Aggregator service
- Safe close all threads for Health Aggregator cleanup
- Added Endpoint GET - /ResourceState, to get the list of resources, their health state, and reason if they are down
- Added Endpoint POST - /Reset, to reset the Health Aggregator service, and initiate fresh connection to all resources
- Added tests for the application