SRE_ScreenCloud_Challenge

SRE Challenge for screenCloud

Prerequisites

Install LibreOffice for Windows here: https://www.libreoffice.org/download/download/

Install Python3, version used at time of writing this "Python 3.8.8", select option Windows installer (64-bit) & select the 'add Python ENV var to PATH' option when installing https://www.python.org/downloads/release/python-397/

Install VisualStudio Code IDE, select Windows Installer https://code.visualstudio.com/download

Checkout remote GitHub repository containing source files to run parser

$: git clone https://github.com/michaels0184/SRE_ScreenCloud_Challenge.git

Run the parser

Open VSCode and start a new Terminal session within editior, once open make sure to change directory to remote repo

$: cd SRE_ScreenCloud_Challenge

Now to run the parser

$: python parse_logfile.py

This will parse and format the information from the logfile "sre-challenge.log" and output results into seperate .csv files within the output dir

output/
  facebook.csv
  instagram.csv
  profle.csv
  profile.csv
  upload-media.csv
  weather.csv

View results

Now open any of the .csv files located within the output dir to find relevant results for a specific path contained within the logs. eg "facebook"

• Combined findings using results generated by the parser (parse_logfile.py)

Path	200 count	403 count	500 count	Total responseTime(ms)	Total no Bytes
Facebook	295	666	30	129051	9839302
Instagram	0	0	570	17130000	5801806
Weather	4328	43	1412	2543298	86127640
Profile	1525	2	3	197031	15240137
Upload-media	275	0	0	2092026	413564759
Profle	745	28	72	451304	8370824

SRE Challenge

Where should engineering focus first to reduce errors? A: The Weather component has the highest count of 500 internal server error message which could be a number of things on the server side.

If performance on this box is slow, which service should we investigate first? A: Instagram, as this service is having the highest total count of combined responseTime and 2nd highest count of 500 errors. Instagram also has 0 recorded successfull 200 logs found which is concerning and perhaps timing out causing the high response time. I would then look at the weather application next as it has the second highest responseTime and the highest 500 error message count.

If we were going to split this application into multiple, what setup might give a good effort/reward ratio? A: Due to the nature of upload-media component, it would be wise to split this into it’s own service potentially on another server as it requires the highest no of combined bytes compared to other components. It might be good also splitting Instagram into it's own app to handle any requests against the Instagram path, as it has the highest count for responseTime and therefore might slow down the overall application.