Experiences with Microservices at Tuenti

This is the talk Aarón Fas and Andrés Viedma presented in the JBcnConf 2015 telling their experiences at Tuenti using a distributed architecture of microservices.

Transcript

1. .Experiences with Microservices at Aarón Fas Andrés Viedma
2. Microservices? I know what you’re probably thinking...
6. Who did you say these guys are? Andrés Viedma @andres_viedma Aarón Fas @aaronfc Java dinosaur Useless gadgets buyer
7. About
8. From Social Network...
9. From Social Network...
10. To Mobile Operator (full MVNO)
11. The PHP Monolith One single source repository PHP???
12. Do you need a release? Take a ticket and wait...
13. .Microservices
14. Microservices… again… (and take a shot) ❖ Distributed, independently deployable components ❖ Well defined interfaces ❖ Simple communication interface (HTTP?) ❖ Each service has its own DB ❖ Each service has its own source repository
15. Microservices… again… (and take a shot) ❖ Distributed, independently deployable components ❖ Well defined interfaces ❖ Simple communication interface (HTTP?) ❖ Each service has its own DB ❖ Each service has its own source repository THAT IS SOA !!!
16. Microservices… again… (and take a shot) ❖ Distributed, independently deployable components ❖ Well defined interfaces ❖ Simple communication interface (HTTP?) ❖ Each service has its own DB ❖ Each service has its own source repository Is that important enough to deserve a new name???
17. Mixing technologies ❖ Allows using different languages ❖ Different platform versions ❖ Incremental technology changes / evolution
18. Separation of responsibilities ❖ Forces separation of responsibilities ➢ Subsystems with well defined facades ➢ Different source repositories
19. Separation of responsibilities ❖ Forces separation of responsibilities ➢ Subsystems with well defined facades ➢ Different source repositories YOU DON’T NEED MICROSERVICES! USE JARS !!!
20. Continuous deployment «Our highest priority is to satisfy the customer through early and continuous delivery of valuable software.» «The best architectures, requirements, and designs emerge from self-organizing teams.» -- Principles of the Agile Manifesto
21. Continuous deployment «Our highest priority is to satisfy the customer through early and continuous delivery of valuable software.» «The best architectures, requirements, and designs emerge from self-organizing teams.» -- Principles of the Agile Manifesto 1 Service => 1 Team? Better than Continuous delivery!: Continuous deployment Team responsible of the deployments?
22. Beware! High costs ❖ No transactions! ➢ Distributed tx? ❖ Requires a much more complex infrastructure ❖ Difficult integration testing
23. For us: Seemed like a good idea ❖ We have small self-organized teams => Continuous deployment is a reality ❖ We wanted Java, we had PHP ❖ Strong SRE / DevOps team ❖ Our software was intended mainly to access 3rd parties => transactions not possible anyway
24. .Communications protocol
25. Existing libraries ❖ No PHP implementation ➢ Avro, Etch, Netflix stack ❖ Only serialization ➢ Protocol buffers ❖ Didn’t exist or were too new ➢ Cap’n Proto, gRPC ❖ Thrift? ➢ Good option, but a lot of PHP boilerplate
26. TService ❖ Own abstraction layer - RPC based ❖ Basic implementation: JSON-RPC ❖ Interface Definition Language (IDL) ❖ Generates Java / PHP / Erlang: ➢ Interchange objects ➢ Client ➢ Server stub
27. TService IDL /** * Manages the transfer of balance between subscriptions. * @version 1 */ interface BalanceTransferService { /** Transfer money from one subscription to another one. */ String transfer(Donation donation) throws NoSuchSubscriptionException; (...) } /** Donation between two subscriptions. */ class Donation { /** Id of the donor */ long from; /** Amount of money to transfer */ int amount; (...) } class NoSuchSubscriptionException extends Exception { int code = 100; } Java???
28. TService Versioning Interface v1 Service Client 1 Client 2 (compatible changes) • New methods • New fields in objects • New parameters in methods • Delete methods / parameters / fields
29. TService Versioning Interface v1 Service Interface v2 Client 1 Client 2 (compatible changes)
30. TService Versioning Interface v1 Service Interface v2 Client 1 Client 2 (compatible changes)
31. .Java Platform
32. Technology stack
33. XConfig ❖ Own configuration system ❖ YAML files based ❖ Git repository ❖ Overriding system: by env, common / service ❖ Hot reloading ➢ Everything adjusts to changes: even DB pools! ➢ No restart required
34. Async jobs TService request processing Enqueue job Queued jobs Executor thread pool
35. Async jobs TService request processing Enqueue job Queued jobs Executor thread pool Cron jobs Cron jobs programming in config
36. Feature disabling ❖ Activation / deactivation of features by config ➢ Is the new development risky? ➢ Is the rest of services / environment ready for the change? ❖ Partial activation of a feature for a % of users ➢ Incremental activation of an optional risky change ➢ A / B tests
37. Integration tests ❖ Custom JUnit runner ➢ Bootstraps the platform ➢ Cleans / restarts the local database ➢ Allows the use of @Inject in tests ➢ Allows overriding in dependency injection => inject mocks of the other services ❖ Uses special, “development” XConfig repo
38. .Monitoring
39. Monitoring, a priority ❖ What is happening or has happened? ➢ Logs ➢ Metrics ➢ Alarms ❖ Distributed architectures are much more difficult to track
40. And basically because...
41. .Let’s talk about logs
42. Logging ❖ Logging library in Java? ➢ Log4j ❖ We needed full details ➢ Filters to expand/simplify information logged ➢ Multiple appenders logged into distinct storages
43. ❖ Overview of appenders Logging log.info(...); Logger MySQL Appender LogStash Appender Hadoop Appender
44. ❖ Following call’s path (TService calls logging) Logging ServiceA ServiceB ServiceC GlobalID = 100 RequestID = 1 GlobalID = 100 RequestID = 2 GlobalID = 100 RequestID = 3 Benefits • Locate in/out for calls • Get all interactions
45. Logging ❖ Kibana dashboard What does it look like?
46. Change query
47. Customize filters
48. Log types by color
49. Full log details
50. .Let’s talk about metrics
51. Metrics ❖ We graphs ➢ As easy as possible to track new metrics ❖ Do not reinvent the wheel ➢ Already using StatsD/Graphite on PHP side ❖ What are we tracking? ➢ Basic monitoring metrics added by the platform ➢ Metrics from Tomcat JMX ➢ Metrics related to business
52. Metrics ❖ Multiple graphs dashboards tested ➢ Default graphite one ➢ Grafana
53. Graphite’s is a little ugly...
54. Grafana is prettier
55. Layout customized
56. Much better UI to create graphs
57. .Let’s talk about alarms
58. Alarms ❖ Graphs are ok, but we don’t have people 24x7 staring at them. ➢ We need notifications ❖ Different things to monitor ➢ SQL queries ➢ Graphite metrics ➢ HTTP requests ➢ ...
59. Alarms ❖ Created our own alarms system ➢ Multiple data sources and easily extensible ➢ Quick edition of conditions ➢ Observers for alarms ❖ We ended up using mainly ➢ MySQL and Graphite data sources ➢ Java Expression Language on config checkers ➢ Email notifications
60. … and then, we found Cabot Separated by service
61. Cabot overview Multiple integrations
62. Cabot overview Service status overview
63. Cabot overview Graphite checks
64. Cabot overview (Creating new check) Set graphite metric
65. Cabot overview (Creating new check) Check data
66. Cabot overview (Creating new check) Set check type/value
67. Cabot overview (Creating new check) Set importance
68. Cabot ❖ Benefits of using Cabot ➢ Friendlier UI than config files ➢ No dependency on the service monitored ➢ Opensource and many integrations
69. Alarms ❖ Where are we heading now? ➢ Moving now most Graphite alarms to Cabot ➢ Replacing thresholds with dynamic expectations (Holt Winters) ❖ It is still the main alarms platform being used
70. .That’s all about monitoring
71. .Some Lessons learned
73. GO ASYNC!!!
74. Don’t get blocked for too long ❖ Concurrent requests: don’t wait for free threads ➢ Own Rate limit mechanism ➢ Tune container thread pool size ➢ Tune database pool (and other possible blocking pools) ❖ Tune clients timeout ➢ It may depend on called service / operation ➢ It may depend on the caller
75. Asynchronous logging log.info(...) Appender MySQL Appender Logstash Appender Hadoop Logger
76. Asynchronous logging log.info(...) When the ring buffer is full… WAIT! Appender MySQL Appender Logstash Appender Hadoop Logger Ring buffer Async Logger Not configurable!
77. Asynchronous logging log.info(...) When the ring buffer is full… WAIT! Appender MySQL Appender Logstash Appender Hadoop Logger Ring buffer Async Logger Async Appender Async Appender Async Appender Not configurable! When async appender full, messages are discarded
78. Asynchronous operations ❖ Getters ➢ Make them fast (sacrifice consistency) ➢ Cache ➢ Use default values ❖ Setters ➢ No operation result ➢ Wait for a notification of operation finished ➢ Query the status of the change
79. Message queues ❖ Operation queues ➢ Retry system ➢ Persistent queues ❖ Publish / subscribe model (pub/sub) ➢ Event driven ➢ Reactive programming
80. Circuit breaker ❖ From the client, consider the status of the service ➢ Previous calls ➢ Health checks ❖ If it’s degraded, don’t call it (close the circuit) ➢ Return a default response ➢ Enqueue the operation for later retry ➢ Throw an error
81. .Do It Yourself
82. Many implementations available ❖ Communication layer ➢ gRPC, Cap’n proto, Thrift… ➢ REST, JSON… ❖ Services platform ➢ Spring boot, Dropwizard, Spark, Ninja, Jodd… ❖ Netflix stack ➢ Hystrix, Ribbon…
83. Make your own combination! (it can’t be so difficult…)
84. Aarón Fas @aaronfc Andrés Viedma @andres_viedma .Thanks! Questions?