CodeCamp23 Enabling Microservice Success

Microservices can be a very effective approach to speeding up delivery of value to your organisation and to your customers. If you get them right.

If you don’t, then microservices are just something that makes everything you do more complicated, from working out where things are going wrong to upgrading the same dependency in tens or hundreds of services.

Sarah will draw on ten years of experience building and operating microservices at the Financial Times to talk about:
• what it means to be “successful” with your architecture and technology choices. Can you move fast? Understand what is happening in production? Avoid having to start again?
• what you need in place to make microservices work, from technology maturity and leadership support through to platform engineering and observability
• where we go next with microservices: what are the new tools and approaches that will help us get the most out of them?

Transcript

1. @sarahjwells@mastodon.social Enabling Microservice Success Sarah Wells
2. @sarahjwells@mastodon.social Enabling Microservice Success Sarah Wells
3. @sarahjwells@mastodon.social
4. @sarahjwells@mastodon.social
5. @sarahjwells@mastodon.social
6. @sarahjwells@mastodon.social What does a successful software development organisation look like?
7. @sarahjwells@mastodon.social
8. @sarahjwells@mastodon.social Accelerate defines performance in terms of the impact on the productivity, profitability and market share of your business, compared to competitors
9. @sarahjwells@mastodon.social Four outcomes that link to high performance: • short lead time • high deployment frequency • low change fail percentage • short mean time to restore
10. @sarahjwells@mastodon.social Accelerate's key capabilities (Forsgren, Humble, Kim)
11. @sarahjwells@mastodon.social • Regularly delivering real business value
12. @sarahjwells@mastodon.social 1. Regularly delivering real business value 2. Adapting to change 3. Maintaining appropriate service levels 4. Spending most of your time on meaningful work 5. Not having to start again 6. Keeping risk to an acceptable level
13. @sarahjwells@mastodon.social Reasons to choose microservices
14. @sarahjwells@mastodon.social Microservices are a solution to an organisational problem
15. @sarahjwells@mastodon.social Reasons to use microservices • Because loose coupling supports flow
16. @sarahjwells@mastodon.social Reasons to use microservices • Because loose coupling supports flow • For the developer experience
17. @sarahjwells@mastodon.social Reasons to use microservices • Because loose coupling supports flow • For the developer experience • Aids compliance and security
18. @sarahjwells@mastodon.social Reasons to use microservices • Because loose coupling supports flow • For the developer experience • Aids compliance and security • Scaling only where needed
19. @sarahjwells@mastodon.social Reasons to use microservices • Because loose coupling supports flow • For the developer experience • Aids compliance and security • Scaling only where needed • Increased robustness
20. @sarahjwells@mastodon.social Reasons to use microservices • Because loose coupling supports flow • For the developer experience • Aids compliance and security • Scaling only where needed • Increased robustness • Increased flexibility
21. @sarahjwells@mastodon.social
22. @sarahjwells@mastodon.social Challenges
23. @sarahjwells@mastodon.social Challenges • Estate complexity
24. @sarahjwells@mastodon.social Challenges • Estate complexity • Operational complexity
25. @sarahjwells@mastodon.social Challenges • Estate complexity • Operational complexity • Finding the right granularity
26. @sarahjwells@mastodon.social Challenges • Estate complexity • Operational complexity • Finding the right granularity • Handling change
27. @sarahjwells@mastodon.social Challenges • Estate complexity • Operational complexity • Finding the right granularity • Handling change • Developer experience is very different
28. @sarahjwells@mastodon.social Challenges • Estate complexity • Operational complexity • Finding the right granularity • Handling change • Developer experience is very different • Organisational change
29. @sarahjwells@mastodon.social Conditions for Success
30. @sarahjwells@mastodon.social Challenges • Domain understanding
31. @sarahjwells@mastodon.social Challenges • Domain understanding • Products not projects
32. @sarahjwells@mastodon.social Challenges • Domain understanding • Products not projects • Leadership support
33. @sarahjwells@mastodon.social Challenges • Domain understanding • Products not projects • Leadership support • Teams that want autonomy
34. @sarahjwells@mastodon.social Challenges • Domain understanding • Products not projects • Leadership support • Teams that want autonomy • Support for that autonomy
35. @sarahjwells@mastodon.social Challenges • Domain understanding • Products not projects • Leadership support • Teams that want autonomy • Support for that autonomy • Technical maturity
36. @sarahjwells@mastodon.social 1. Regularly delivering real business value 2. Adapting to change 3. Maintaining appropriate service levels 4. Spending most of your time on meaningful work 5. Not having to start again 6. Keeping risk to an acceptable level
37. @sarahjwells@mastodon.social 1. Regularly delivering real business value
38. @sarahjwells@mastodon.social Four outcomes that link to high performance: • short lead time • high deployment frequency • low change fail percentage • short mean time to restore
39. @sarahjwells@mastodon.social
40. @sarahjwells@mastodon.social Accelerate's key capabilities (Forsgren, Humble, Kim)
41. @sarahjwells@mastodon.social 2. Adapting to change
42. @sarahjwells@mastodon.social Changing business priorities
43. @sarahjwells@mastodon.social
44. @sarahjwells@mastodon.social Changing technology
45. @sarahjwells@mastodon.social
46. @sarahjwells@mastodon.social "For many years, the tongue-in-cheek definition of software architecture was “the stuff that’s hard to change later.” Later, the microservices architecture style appeared, where change is a first-class design consideration." From the preface of Fundamentals of Software Architecture, Mark Richards and Neal Ford
47. @sarahjwells@mastodon.social Beware the "long tail"
48. @sarahjwells@mastodon.social 3. Maintaining appropriate service levels
49. @sarahjwells@mastodon.social Four outcomes that link to high performance: • short lead time • high deployment frequency • low change fail percentage • short mean time to restore
50. @sarahjwells@mastodon.social Optimising for a short mean time to restore
51. @sarahjwells@mastodon.social Loose coupling = small blast radius
52. @sarahjwells@mastodon.social https://lamport.azurewebsites.net/pubs/distributed-system.txt “A distributed system is one in which the failure of a computer you didn't even know existed can render your own computer unusable”
53. @sarahjwells@mastodon.social Richard I Cook - https://www.adaptivecapacitylabs.com/ HowComplexSystemsFail.pdf
54. @sarahjwells@mastodon.social Richard I Cook - https://www.adaptivecapacitylabs.com/ HowComplexSystemsFail.pdf Distributed systems need to be built with multiple layers of defense against failure
55. @sarahjwells@mastodon.social Build resilience in
56. @sarahjwells@mastodon.social
57. @sarahjwells@mastodon.social
58. @sarahjwells@mastodon.social Know when something important is broken
59. @sarahjwells@mastodon.social Checks running in January 2021 at the Financial Times
60. @sarahjwells@mastodon.social
61. @sarahjwells@mastodon.social
62. @sarahjwells@mastodon.social Focus on business capabilities
63. @sarahjwells@mastodon.social Monitor your business capabilities
64. @sarahjwells@mastodon.social Work out what's going on
65. @sarahjwells@mastodon.social Monitoring is not enough
66. @sarahjwells@mastodon.social Observability: can you infer what’s going on in the system by looking at external outputs?
67. @sarahjwells@mastodon.social Tracing events through your system
68. @sarahjwells@mastodon.social
69. @sarahjwells@mastodon.social 4. Spending most of your time on meaningful work
70. @sarahjwells@mastodon.social Autonomous teams have a lot of things to know about
71. @sarahjwells@mastodon.social Maintaining microservices means constant work
72. @sarahjwells@mastodon.social Product development teams can’t be completely autonomous
73. @sarahjwells@mastodon.social Platform engineering
74. @sarahjwells@mastodon.social Platform engineering Engineering Enablement
75. @sarahjwells@mastodon.social Paving the road
76. @sarahjwells@mastodon.social https://charity.wtf/2018/12/02/software-sprawl-the-golden-p ath-and-scaling-teams-with-agency/
77. @sarahjwells@mastodon.social • Self service • Well-documented • Discoverable
78. @sarahjwells@mastodon.social 5. Not having to start again
79. @sarahjwells@mastodon.social Without work, systems get messy
80. @sarahjwells@mastodon.social Starting again is appealing!
81. @sarahjwells@mastodon.social Rebuilding stops you delivering value
82. @sarahjwells@mastodon.social • £10 million • 2 years in maintenance mode
83. @sarahjwells@mastodon.social Microservices *should* make it easy
84. @sarahjwells@mastodon.social "Habitability is the characteristic of source code that enables people coming to the code later in its life to understand its construction and intentions and to change it comfortably and confidently." Habitability and Piecemeal Growth, Richard P Gabriel, https://www.dreamsongs.com/Files/PatternsOfSoftware.pdf
85. @sarahjwells@mastodon.social 6. Keeping risk to an acceptable level
86. @sarahjwells@mastodon.social Know your estate
87. @sarahjwells@mastodon.social https://www.infoq.com/presentations/productivity-ft/
88. @sarahjwells@mastodon.social Active ownership
89. @sarahjwells@mastodon.social Services need to be actively owned, by a team
90. @sarahjwells@mastodon.social
91. @sarahjwells@mastodon.social Be flexible on tech choices only where you need to be
92. @sarahjwells@mastodon.social The paved road allows you to make global decisions around risk
93. @sarahjwells@mastodon.social • Safe • Secure • Reliable by default
94. @sarahjwells@mastodon.social Guardrails and governance
95. @sarahjwells@mastodon.social The FT’s guardrails
96. @sarahjwells@mastodon.social
97. @sarahjwells@mastodon.social log4shell
98. @sarahjwells@mastodon.social What next for microservices? https://thenewstack.io/the-future-of-microservices-more-abstr actions/
99. @sarahjwells@mastodon.social https://thenewstack.io/the-future-of-microservices-more-abstr actions/ We have mature well-designed microservice frameworks
100. @sarahjwells@mastodon.social https://thenewstack.io/the-future-of-microservices-more-abstr actions/ Open telemetry provides a standard for distributed telemetry data
101. @sarahjwells@mastodon.social Improving developer experience is the next challenge
102. @sarahjwells@mastodon.social https://thenewstack.io/the-future-of-microservices-more-abstr actions/
103. @sarahjwells@mastodon.social Daniel Bryant, quoted in https://thenewstack.io/ the-future-of-microservices-more-abstractions/ “companies.. are creating a Heroku-like CLI for their developers, so that a command like ‘create new microservice’ spins up some scaffolding, plugs into CI, plugs into pipelines, plugs into observability”
104. @sarahjwells@mastodon.social https://backstage.io/
105. @sarahjwells@mastodon.social https://redmonk.com/jgovernor/2022/02/21/what-is-developer -experience-a-roundup-of-links-and-goodness/ “organisations want developers to spend more time coding and less time thinking about and working on infrastructure”
106. @sarahjwells@mastodon.social https://redmonk.com/jgovernor/2022/02/21/what-is-developer -experience-a-roundup-of-links-and-goodness/ “Why would a developer put up with a poor developer experience when they can likely find a job that better aligns with their needs and preferences?”
107. @sarahjwells@mastodon.social Wrapping up…
108. @sarahjwells@mastodon.social Fast flow is important
109. @sarahjwells@mastodon.social Which means you need: • zero-downtime deployments • loose coupling
110. @sarahjwells@mastodon.social Microservices reduce build & deploy complexity and increase operational complexity
111. @sarahjwells@mastodon.social Autonomy comes with responsibility Jason Yip, Thoughts on the Spotify model, https://jchyip.medium.com/my- critique-of-the-spotify-model-part-1-197d335ef7af
112. @sarahjwells@mastodon.social You need some scaffolding
113. @sarahjwells@mastodon.social Everything is a bit distributed now
114. @sarahjwells@mastodon.social https://www.oreilly.com/library/view/enabling-microservice-su ccess/9781098130787/
115. @sarahjwells@mastodon.social Slides: https://speakerdeck/sarahwells/code-camp23- enabling-microservice-success Contact: @sarahwells@mastodon.social https://www.linkedin.com/in/sarahjwells1/