Who What How: Understanding Kubernetes Development through DevStats - Josh Berkus & Łukasz Gryglicki
summary
DevStats pulls GitHub Archive events plus Git and limited GitHub API calls into Postgres, rolls hourly metrics into InfluxDB and Grafana at k8s.devstats.cncf.io, and exposes charts like PR time to engagement (15th/50th/85th percentiles) and PR workload per SIG that multiplies PR count by size and divides by active reviewers. It shows bot-command trends (e.g. slash retest spikes and /approve surges at releases) and helped detect overloaded SIGs (Node, API machinery) so reviewers were reassigned during 1.9.
AI-written summary · about
chapters (8)
- 0:00Start
- 1:19I've got to know!
- 4:21GitHub stats limitations
- 14:03Open and Close (7day moving average)
- 22:31Data flow: big picture
- 23:09Sources: Github Archives
- 27:57Databases
- 29:44Steps for filing a request
related talks
35mRegulations, Ethics And Messy Data: Tales From An Open Source Analyst
30mData Tales from an Open Source Research Team
13mShaping Kubernetes Community Culture
31mKubernetes SIG Contributor Experience: A Look Back, a Step Forward
27mCommunity Data is Not Community Metrics
19mOSPO Data Pyramid
01h 04mWhat can we learn from 750 billion GitHub events and 42 TB of code
24mA Beginners Guide to Open Source Metrics and Analysis