StackEngine Demo - Docker Austin
3 likes6,089 views
The document outlines the challenges and developments in managing Docker applications in production, emphasizing the importance of DevOps culture and practices in enhancing business outcomes. It discusses issues such as container sprawl, deployment difficulties, and the need for centralized control and telemetry to improve operational efficiency. The creators of StackEngine aim to address these challenges by providing solutions that enhance developer productivity and streamline application deployment in containerized environments.
StackEngine Demo - Docker Austin
- 1. Managing Docker Applications in Production Boyd E. Hemphill - @behemphi Mike Raab - @stackengine
- 2. History - Boyd Oracle Developer, MySQL DBA, PHP Developer, System Architect, DevOps Director (snickering encouraged), Evangelist
- 3. History - Boyd Oracle Developer, MySQL DBA, PHP Developer, System Architect, DevOps Director (snickering encouraged) Started Austin DevOps In 2012
- 4. History - Boyd Oracle Developer, MySQL DBA, PHP Developer, System Architect, DevOps Director (snickering encouraged) Started Austin DevOps In 2012 Co-Organizer of DevOps Days Austin
- 5. History - Boyd Oracle Developer, MySQL DBA, PHP Developer, System Architect, DevOps Director (snickering encouraged) Started Austin DevOps In 2012 Co-Organizer of DevOps Days Austin Became involved in Docker Austin in August.
- 6. History - Boyd Oracle Developer, MySQL DBA, PHP Developer, System Architect, DevOps Director (snickering encouraged) Started Austin DevOps In 2012 Co-Organizer of DevOps Days Austin Became involved in Docker Austin in August. Working on the first Container Days Austin for late Feb
- 7. History - Boyd Oracle Developer, MySQL DBA, PHP Developer, System Architect, DevOps Director (snickering encouraged) Started Austin DevOps In 2012 Co-Organizer of DevOps Days Austin Became involved in Docker Austin in August. Working with Docker for the first Docker Days Austin Keep an eye on @stackengine to see if we can make it happen!
- 8. History - StackEngine Started in May 2014 to address challenges facing Docker adopters
- 9. History - StackEngine Started in May 2014 to address challenges facing Docker adopters Funding announcement October
- 10. History - StackEngine Started in May 2014 to address challenges facing Docker adopters Funding announcement October Currently in private beta
- 11. History - StackEngine Started in May 2014 to address challenges facing Docker adopters Funding announcement October Currently in private beta mike@stackengine.com
- 12. Common Ground
- 13. What is The Goal of Your Company?
- 14. What is The Goal of Your Company? Make Money!
- 15. So … What is DevOps?
- 16. Is DevOps a Process?
- 17. Is it an intersection of overlapping concerns?
- 18. Is DevOps a Culture?
- 19. DevOps: DevOps is the way in which a technology organization embeds itself in a business to the benefit of that business.
- 20. Presentation Prelude Describe the problems StackEngine sees in using containers in production.
- 22. Netflix (unicorn) Service Oriented Architecture
- 23. Benefits Feature velocity - Increase developer productivity using containers to model application topology
- 24. Benefits Feature velocity - Increase developer productivity using containers to model application topology Innovation velocity - Decreased build and test times make more time for trying that next great idea!
- 25. Benefits Feature velocity - Increase developer productivity using containers to model application topology Innovation velocity - Decreased build and test times make more time for trying that next great idea! Density - Reduce Costs
- 26. Benefits Feature velocity - Increase developer productivity using containers to model application topology Innovation velocity - Decreased build and test times make more time for trying that next great idea! Density - Reduce Costs Separation of Concerns - “Code to an interface” moves to the infrastructure level
- 27. Benefits Taken together a company that can get to the end state has a significant and lasting competitive advantage
- 29. Concerns This is still very early days for Docker
- 30. Concerns This is still very early days for Docker It is even earlier for Rocket
- 31. Concerns This is still very early days for Docker It is even earlier for Rocket Tools to support containerized applications are immature or do not exist
- 32. Concerns This is still very early days for Docker It is even earlier for Rocket Tools to support containerized applications are immature or do not exist Service oriented architecture is hard
- 33. Concerns This is still very early days for Docker It is even earlier for Rocket Tools to support containerized applications are immature or do not exist Service oriented architecture is hard Increasing developer throughput is only a local optima, thus a false economy
- 34. Concerns Docker adoption is the first major DevOps initiative not driven by Ops.
- 35. Concerns Docker adoption is the first major DevOps initiative not driven by Ops. Laundry Ops: http:// stackengine.com/ laundryops- practical-devops- at-home/
- 36. SDLC Developers are moving forward
- 37. SDLC Developers are moving forward QA sees production modeling, large test grids, more effective integration testing
- 38. SDLC Developers are moving forward QA sees production modeling, large test grids, more effective integration testing Build engineers see parallelized build grids for faster feedback and better build quality
- 39. xkcd … er … SDLC Operations panics - no tools or instrumentation
- 40. SysAdmin Don’t want this guy panicking
- 42. Information Architecture Problem Description StackEngine’s work on that problem today
- 43. Information Architecture Problem Description StackEngine’s work on that problem today StackEngine’s planned work on that problem for the future
- 44. Problem #1 The Final Mile
- 45. The Final Mile Developer uses containers
- 46. The Final Mile Developer uses containers System is built and images are produced
- 47. The Final Mile Developer uses containers System is built and images are produced Images stored in a private registry
- 48. The Final Mile Developer uses containers System is built and images are produced Images stored in a private registry Now we need to deploy to a QA environment
- 49. The Final Mile So we build a manual solution depending on human behavior
- 50. The Final Mile So we build a manual solution depending on human behavior Someone builds a tool
- 51. Crossing the Final Mile - Today A Simple Solution to Deployment
- 52. Deploy My Stuff - Today Create a component to represent your registry
- 53. Deploy My Stuff - Today Create a component to represent the image you would like to run
- 54. Deploy My Stuff - Today Distribute that image and run it on the hosts in your Mesh.
- 55. Deploy My Stuff - Today Magic happens.
- 56. Deploy My Stuff - Today Life is good
- 57. Crossing the Final Mile - Future Heterogeneous, Policy Driven Deployment
- 58. Deploy My Stuff - Future Support other container types
- 59. Deploy My Stuff - Future Support other container types Policy Based Deployment
- 60. Deploy My Stuff - Future We should put another bunny cloud over there …
- 63. Container Sprawl First there was VMware and suddenly a large corps departments could easily spin up environments
- 64. Container Sprawl First there was VMware and suddenly a large corps departments could easily spin up environments Then there was the cloud and developers could completely circumvent ops
- 65. Container Sprawl First there was VMware and suddenly a large corps departments could easily spin up environments Then there was the cloud and developers could complete circumvent Ops Now there is Docker and Developers are creating numerous abstractions that are not managed like traditional VMs
- 66. Container Sprawl First there was VMware and suddenly a large corps departments could easily spin up environments Then there was the cloud and developers could complete circumvent Ops Now there is Docker and Developers are creating numerous abstractions that are not managed like traditional VMs Rocket potentially segments the Market and means different performance, security, runtime and management characteristics
- 68. Container Sprawl Containers, Dockerfiles, and Images on VMs VMs on physical hardware located in availability zones
- 69. Container Sprawl Containers, Dockerfiles, and Images on VMs VMs on physical hardware hardware located in availability zones AZs in data centers in geographic regions.
- 70. My Head Hurts Container Sprawl
- 71. Artifact Inventory - Today A drillable and searchable view of infrastructure
- 72. Where’s My Stuff - Today Hosts
- 73. Where’s My Stuff - Today Images
- 74. Where’s My Stuff - Today Containers
- 75. Where’s My Stuff - Today Filters and Build Actions
- 76. Where’s My Stuff - Today Search
- 77. Artifact Inventory - Future Geographically aware physical topology
- 78. Where’s My Stuff - Future Geographically drillable physical topologies
- 79. Where’s My Stuff - Future Geographically drillable physical topologies Policy based inventory levels (pull not push)
- 80. Where’s My Stuff - Future Hey, why are there two purple stripes?
- 83. Telemetry What does monitoring a containerized app look like?
- 84. Telemetry What does monitoring a containerized app look like? http://goo.gl/VkAQo4
- 85. Telemetry What does monitoring a containerized app look like? http://goo.gl/VkAQo4 What does log aggregation look like?
- 86. Telemetry What does monitoring a containerized app look like? http://goo.gl/VkAQo4 What does log aggregation look like? http://goo.gl/uNPe9P
- 87. Decision Dimensions - Today Host level metrics to help you manage running containers.
- 88. What’s my stuff doing - Today I have this much stuff.
- 89. What’s my stuff doing - Today My hosts are not yet stuffed.
- 90. What’s my stuff doing - Today The stuffing in a host is as expected.
- 91. What’s my stuff doing - Today A bit of the stuffing is doing what it should.
- 92. What’s my stuff doing - Today What is the stuff in this image?
- 93. What’s my stuff doing - Today How do I visualize the stuff in an image?
- 94. Decision Dimensions - Future There is never too much information about run time.
- 95. What’s My Stuff Doing - Future Improve Host Metrics
- 96. What’s My Stuff Doing - Future Improve Host Metrics Improve Image Metrics
- 97. What’s My Stuff Doing - Future Improve Host Metrics Improve Image Metrics Add Container Metrics
- 98. What’s My Stuff Doing - Future Ahh … look … the bunny on the left is happy!
- 100. Problem #4 Centralized Control Plane
- 101. Centralized Control Plane The ability to act upon a system from a single point of entry
- 102. Centralized Control Plane The ability to act upon a system from a single point of entry Actions should happen next to the information causing the action
- 103. Centralized Control Plane The ability to act upon a system from a single point of entry Actions should happen next to the information causing the action Actions should happen within the context of inventory
- 104. Centralized Control Plane The ability to act upon a system from a single point of entry Actions should happen next to the information causing the action Actions should happen within the context of inventory UI, API, CLI
- 105. Actions - Today Preventative, automatic, deterministic
- 106. Make Stuff do Stuff - Today Manage Images
- 107. Make Stuff do Stuff - Today Manage Containers
- 108. Actions - Future Initialization, Taxonomy, Corrective
- 109. Make Stuff do Stuff - Future Policy based actions by the Controller
- 110. Make Stuff do Stuff - Future Policy based actions by the Controller Powerful Control Plane Taxonomy (Tagging)
- 111. Make Stuff do Stuff - Future Policy based actions by the Controller Powerful Control Plane Taxonomy (Tagging) Ants not Cattle
- 112. Make Stuff do Stuff - Future Blow really hard and make the bunnies kiss.
- 115. Heterogeneous Service Providers Containers live on services (IaaS, PaaS)
- 116. Heterogeneous Service Providers Containers live on services (IaaS, PaaS) Images come from different places (Docker Hub, Quay.io)
- 117. Heterogeneous Service Providers Containers live on services (IaaS, PaaS) Images come from different places (Docker Hub, Quay.io) Within an IaaS service provider are other services (Databases, load balancers)
- 118. Heterogeneous Service Providers Containers live on services (IaaS, PaaS) Images come from different places (Docker Hub, Quay.io) Within an IaaS service provider are other services (Databases, load balancers) Persistent Data stores present their own unique problem as a service
- 119. Components - Today Abstractions for Control Plane Operation
- 120. Building Apps with My Stuff Components are infrastructure service abstractions
- 121. Building Apps with My Stuff Components can be arranged in logical ways to represent physical infrastructure.
- 122. Building Apps with My Stuff Start the db
- 123. Building Apps with My Stuff Start the web server
- 124. Building Apps with My Stuff Run the application server
- 125. Building Apps with My Stuff Run the Cache
- 126. Components - Future More and varied components
- 127. Building Apps with My Stuff - Future More component types is like having more varied Lego bricks
- 128. Building Apps with My Stuff - Future More component types is like having more varied Lego bricks Pluggable component architecture
- 129. Building Apps with My Stuff What else could we make with bunnies and rainbows?
- 131. Roadmap
- 133. Service Discovery Allows multiple deployments of the same service on the same host.
- 134. Service Discovery Allows multiple deployments of the same service on the same host. Containers register themselves upon startup and are found by tags rather than URI’s, fixed ports, etc.
- 135. Service Discovery Allows multiple deployments of the same service on the same host. Containers register themselves upon startup and are found by tags rather than URI’s, fixed ports, etc. We favor something like consul with a search interface like the chef server.
- 136. Service Discovery Allows multiple deployments of the same service on the same host. Containers register themselves upon startup and are found by tags rather than URI’s, fixed ports, etc. We favor something like consul with a search interface like the chef server. A step towards cattle containers instead of pet containers
- 137. Service Discovery Allows multiple deployments of the same service on the same host. Containers register themselves upon startup and are found by tags rather than URI’s, fixed ports, etc. We favor something like consul with a search interface like the chef server. A step towards cattle containers instead of pet containers Ultimate goal is “ants not cattle”
- 139. Roadmap User Management Addition of new Components
- 140. Roadmap User Management Addition of new Components Better telemetry
- 141. Roadmap User Management Addition of new Components Better telemetry Scheduling
- 142. Roadmap User Management Addition of new Components Better telemetry Scheduling A/B Deployment support
- 143. StackEngine is the Stuff It’s easier to dream with StackEngine.
- 144. Live Product Demo Al Gore and his gremlins willing
- 145. Boyd E. Hemphill - @behemphi Mike Raab - @stackengine Thank You for Your Time and Comments.