This project is a solution to the Site Reliability Engineering take-home exercise provided by Fetch. It monitors the availability of HTTP endpoints defined in a YAML file and logs their cumulative availability every 15 seconds. The program ensures that endpoints are only considered available if they respond with a 2xx status code and within 500ms.

• Accepts a YAML configuration file as input
• Makes HTTP requests using optional method, headers, and body
• Normalizes domains (ignores port numbers) to group availability stats
• Checks all endpoints concurrently every 15 seconds
• Tracks cumulative availability per domain
• Drops decimal places from availability percentages as per requirements
• Logs availability in a readable format to stdout

• Python 3.8+
• requests module
• PyYAML

Install dependencies using:

pip install -r requirements.txt

• The script loads endpoint definitions from a YAML file passed as a command-line argument.
• Every 15 seconds, it sends HTTP requests to all endpoints concurrently.
• It tracks whether each response meets the criteria (2xx status code and ≤ 500ms).
• It groups results by domain (ignoring ports) and prints cumulative availability for each.
• Availability is printed as an integer percentage (decimals are dropped).

1. Added a 0.5-second timeout to each HTTP request to ensure quick responsiveness.
2. Measured response duration using timestamps to validate that the endpoint responded within the required 500ms.
3. Implemented a default HTTP method of "GET" in case it is not specified in the YAML configuration.
4. Removed port numbers when determining the domain by using Python's urllib.parse to normalize the hostname.
5. Replaced the original sequential request logic with multithreaded execution so that all endpoints are checked concurrently and efficiently.
6. Used a thread lock (threading.Lock) to prevent race conditions when updating the shared domain_stats dictionary.
7. Updated the availability percentage calculation to use int() instead of round(), so decimal values are dropped, as per the assignment requirements.
8. Changed the request body handling from json= to data= to properly send raw JSON string bodies.
9. Improved the log output format by adding clear headers and separators to distinguish between check cycles.

To meet the requirement of dropping decimal places in the availability percentage:

# Drop decimals in availability percentage as required by the Fetch instructions
availability = int((stats["up"] / stats["total"]) * 100)