|1st place:||300 EUR|
Another day in the office...
Internet connection found. Hot coffee on the table. Programs run. Relaxed mood is here. And suddenly, red light alarm. Mood freezes. Crash detected!
For sure everyone knows this kind of situation. And for sure everyone would like to predict them. In this challenge you can do so (and shape the future)! :)
Can You Help Alfred To Predict A Problem?
Alfred and his team are responsible for enterprise job scheduling in a complex environment. Thousands of jobs are running simultaneously on different environments. Sometimes, jobs are crashing due to unforeseen circumstances. Based on historically run data we want to recognize the pattern why jobs are crashing and want to predict if we are close to a future crash or not.
Your task is to create an algorithm which analyzes historically run data and gives us an indication whether the current situation is similar to a crash pattern in the past. Additionally, try to visualize the critical path in the job scheduling. You will get sufficient processing data which contain all necessary information (start, stop, failure code, predecessor, successor…)
Focus on the following:
Identify what led to a crash
Think how to create an algorithm out of the findings
Think how to visualize the critical path
Formal explanation how to present the critical path for different timeframe (do not exceed one A4 page text).
Formal explanation how to analyze crash situations and how to predict the probability of crash in the future (do not exceed two A4 page text).
Clarity of argumentation
File with all relevant data will be sent to all interested users on daily basis.