Data Science in DevOps/SysOps
summary
Collects CI/CD metadata and builds concrete models: a decision tree classifier predicts when to delete stale containers (it blocks a covert Bitcoin miner), CI test metadata reveals one slow host whose kernel config is fixed, and adaptive timeouts suspend long tests and resume them from container snapshots to free resources. Uses TF-IDF on logs plus keyword models in Python and Elasticsearch to predict failure causes so triage takes minutes instead of days. Practical fixes and measurable wins, not vague buzzwords.
AI-written summary · about
related talks
31mTowards an open operations data community
51mFrom DevOps for data to DataOps
24mUncovering Community and Project Insights Through Data Driven Methods
26mApplying DevOps practices in Data and ML Engineering
28mDevOps in a Data Warehouse: Inside Out
26mThe Hidden Foundation: How Effective Data Management Determines AI System Success
13mPredictive Insights for IT Operations #bbuzz
01h 17mData Engineering in the AI Era: Building a Robust Foundation for AI Applications