Pipelining quality best practices, lessons learned and tools for - PowerPoint PPT Presentation

Pipelining quality best practices, lessons learned and tools for continuous delivery processes 25. Oktober 2017, Markus Tiede – Basler Versicherungen AG Wir machen Sie sicherer. Wir machen Sie sicherer.

@MarkusTiede Wir machen Sie sicherer.

Bird's eye view http://www.datacenterjournal.com/wp-content/uploads/2016/11/devops111116.png 25.10.17 www.baloise.ch 3

01 Galileo „ OOTB approach n large-scale insurance enterprise project n 4 products CM Portal n 8 Scrum teams for "customization" n ± 75 people n POs, SMs, Architects, Developer, Tester, Business Analysts n 1-2 team(s) per product n cross-functional team(s): integration, release www.baloise.ch 25.10.17 4

02 Overall process „ "classic" SCRUM n two week sprints (sync) n but: harmonized releases n decoupled from sprint(s) n everything digital n knowledge / documentation base n deeply integrated / coupled n application lifecycle n CRs, User Stories, Bugs, … n SourceCodeManagement n Continuous Integration www.baloise.ch 25.10.17 5

03 A journey to production „ knowledge and automation is key: Confluence Dashboard www.baloise.ch 25.10.17 6

04 Jenkins pipelines ( scripted and declarative ) „ orchestration of environment(s) – CI as code (Job DSL) LB / SSO PC BC CM Portal o1 o1 o1 o1 o2 o2 o2 o2 coo coo coo o3 www.baloise.ch 25.10.17 7

05 Agile testing pyramid (per coordinate) „ goal: fail fast / fail early www.baloise.ch 25.10.17 8

LB / SSO PC BC CM Portal 06 Coordinates (stack : env) o1 o1 o1 o1 „ allow (massive) parallism + bubbling o2 o2 o2 o2 coo coo coo o3 TEST INT ACC PROD Project+0 TEST INT ACC Project+1 TEST INT ACC Project+2 TEST INT Project+3 12 x 12 = 144 individual nodes www.baloise.ch 25.10.17 9

TEST INT ACC PROD Project+0 07 Releases TEST INT ACC Project+1 „ different flavours TEST INT ACC Project+2 TEST INT n MAJ.MIN.MIC.FIX Project+3 n HOTFIX ( e.g. 4.0.1.2 ) n 8 – 12h (full regression) n MICRO n < 1 week n MINOR n < 1 month n MAJOR n < 1 year www.baloise.ch 25.10.17 10

08 Change and data flow „ multi-master branches and tags per repo 2.3.2.0.rc1 2.4.0.0.rc2 2.4.0.0.rc1 www.baloise.ch 25.10.17 11

09 Change and data flow „ "Golden master / copy"-approach Project+0 TEST INT ACC PROD Project+1 TEST INT ACC Project+2 TEST INT ACC Project+3 TEST INT www.baloise.ch 25.10.17 12

10 Tools „ inspection, analysis, monitoring n low level – log mining n high level – "pure path" analysis n custom level – knowledge graph www.baloise.ch 25.10.17 13

11 2do „ aka – pain points n Environment as code n Puppet, Anisble & Co n Code Review per Feature Branch n automated and manual n (Test) Data Management n all or nothing – x TB of data n Highly interconnected data n Legacy systems / processes (Batches) www.baloise.ch 25.10.17 14

12 Summary „ take away n 100% digital – linked resources; knowledge base n simultanuous release (train) vs. agile shippable artifact n Multi-(Stack)-(Jenkins)-Pipelines n Agile Test-Pyramid n SCM „tag“ driven deployments n Data Management: Golden Data / Master approach n Build your own knowledge graph: neo4j www.baloise.ch 25.10.17 15

https://github.com/baloise Making you safer.