Orange Dev Days 2019

Transcript

1. 1 @loicmdivad @XebiaFr @loicmdivad @XebiaFr Streaming Apps and Poison Pills: handle the unexpected with Kafka Streams 13 Nov. 2019 - Orange Dev Test Day
2. 2 @loicmdivad @XebiaFr @loicmdivad @XebiaFr Loïc DIVAD Data Engineer @XebiaFr @loicmdivad
3. 3 @loicmdivad @XebiaFr Processor API: The dark side of Kafka Streams
4. 4 @loicmdivad @XebiaFr
5. 5 @loicmdivad @XebiaFr La XebiCon‘19 en quelques chiffres 28 novembre 60 conférences techniques 15 retours d'expérience clients 1 500 participants le rendez-vous des C-Levels
6. 6 @loicmdivad @XebiaFr The Rise of the Stream Data Platform
7. 7 @loicmdivad @XebiaFr The Rise of the Stream Data Platform
8. 8 @loicmdivad @XebiaFr The Rise of the Stream Data Platform
9. 9 @loicmdivad @XebiaFr Apache Kafka: Partitions, Messages and Offsets 0 1 2 3 4 5 6 7 8 9 10 11 12 13 CONSUMER A CONSUMER B New Writes
10. 10 @loicmdivad @XebiaFr 10 @loicmdivad @XebiaFr Incoming records may be corrupted, or cannot be handled by the serializer / deserializer. These records are referred to as “poison pills” 1. Log and Crash 2. Skip the Corrupted 3. Sentinel Value Pattern 4. Dead Letter Queue Pattern
11. 11 @loicmdivad @XebiaFr Apache Kafka: Brokers and Clients Kafka Broker Kafka Producer ✉ ✉ Client responsibilities: - batching - serialization - compression ✉ ✉ ✉ ✉ ✉ ✉ ✉ ✉ ✉ ✉ ✉ ✉ ✉ ✉ ✉ Topic Segments Open Segments
12. 12 @loicmdivad @XebiaFr Ratatouille app, a delicious use case Streaming APP
13. 13 @loicmdivad @XebiaFr Ratatouille app, a delicious use case Streaming APP
14. 14 @loicmdivad @XebiaFr 14 @loicmdivad @XebiaFr Streaming App Poison Pills 1. Log and Crash - Breakfast 2. Skip the Corrupted - Lunch 3. Sentinel Value Pattern - Drink 4. Dead Letter Queue Pattern - Dinner
15. 15 @loicmdivad @XebiaFr
16. 16 @loicmdivad @XebiaFr Really old systems receive raw bytes directly from message queues 10100110111010101 Exercise #1 - breakfast
17. 17 @loicmdivad @XebiaFr Really old systems receive raw bytes directly from message queues With Kafka (Connect and Streams) we’d like to continuously transform these messages 10100110111010101 Kafka Connect Kafka Brokers Exercise #1 - breakfast
18. 18 @loicmdivad @XebiaFr Really old systems receive raw bytes directly from message queues With Kafka (Connect and Streams) we’d like to continuously transform these messages But we need a deserializer with special decoder to understand each event What happens if we get a buggy implementation of the deserializer? 10100110111010101 Kafka Connect Kafka Brokers Kafka Streams Exercise #1 - breakfast
19. 19 @loicmdivad @XebiaFr The Tooling Team They will provide an appropriate deserializer
20. 20 @loicmdivad @XebiaFr
21. 21 @loicmdivad @XebiaFr
22. 22 @loicmdivad @XebiaFr
23. 23 @loicmdivad @XebiaFr Take Away
24. 24 @loicmdivad @XebiaFr @loicmdivad @XebiaFr
25. 25 @loicmdivad @XebiaFr Log and Crash
26. 26 @loicmdivad @XebiaFr Log and Crash
27. 27 @loicmdivad @XebiaFr
28. 28 @loicmdivad @XebiaFr
29. 29 @loicmdivad @XebiaFr
30. 30 @loicmdivad @XebiaFr ▼ Change consumer group ▼ Manually update my offsets ▼ Reset my streaming app and set my auto reset to ▽ ▼ Destroy the topic, no message = no poison pill ▽ ▼ My favourite <3 ▽ Don’t Do ▼ Fill an issue and suggest a fix to the tooling team
31. 31 @loicmdivad @XebiaFr @loicmdivad @XebiaFr
32. 32 @loicmdivad @XebiaFr 32 @loicmdivad @XebiaFr Log and Crash Like all consumers, Kafka Streams applications deserialize messages from the broker. The deserialization process can fail. It raises an exception that cannot be caught by our code. Buggy deserializers have to be fixed before the application restarts, by default ...
33. 33 @loicmdivad @XebiaFr
34. 34 @loicmdivad @XebiaFr Stream Topology Streaming APP
35. 35 @loicmdivad @XebiaFr
36. 36 @loicmdivad @XebiaFr • starter • main • dessert
37. 37 @loicmdivad @XebiaFr @loicmdivad @XebiaFr
38. 38 @loicmdivad @XebiaFr Skip the Corrupted
39. 39 @loicmdivad @XebiaFr 39 @loicmdivad @XebiaFr
40. 40 @loicmdivad @XebiaFr
41. 41 @loicmdivad @XebiaFr Take Away
42. 42 @loicmdivad @XebiaFr @loicmdivad @XebiaFr
43. 43 @loicmdivad @XebiaFr 43 @loicmdivad @XebiaFr The Exception Handler in the call stack Powered by the Flow intelliJ plugin ➞ findtheflow.io
44. 44 @loicmdivad @XebiaFr 44 @loicmdivad @XebiaFr Powered by the Flow intelliJ plugin ➞ findtheflow.io The Exception Handler in the call stack
45. 45 @loicmdivad @XebiaFr 45 @loicmdivad @XebiaFr Powered by the Flow intelliJ plugin ➞ findtheflow.io The Exception Handler in the call stack
46. 46 @loicmdivad @XebiaFr 46 @loicmdivad @XebiaFr Powered by the Flow intelliJ plugin ➞ findtheflow.io The Exception Handler in the call stack
47. 47 @loicmdivad @XebiaFr 47 @loicmdivad @XebiaFr Skip the Corrupted All exceptions thrown by deserializers are caught by a A handler returns Fail or Continue You can implement your own Handler But the two handlers provided by the library are really basic… let’s explore other methods
48. 48 @loicmdivad @XebiaFr 48 @loicmdivad @XebiaFr All exceptions thrown by deserializers are caught by a A handler returns Fail or Continue You can implement your own Handler But the two handlers provided by the library are really basic… let’s explore other methods Skip the Corrupted Take Away
49. 49 @loicmdivad @XebiaFr
50. 50 @loicmdivad @XebiaFr Stream Topology Streaming APP
51. 51 @loicmdivad @XebiaFr
52. 52 @loicmdivad @XebiaFr • wine • rhum • beer • champagne • ...
53. 53 @loicmdivad @XebiaFr We need to turn the deserialization process into a pure transformation that cannot crash To do so, we will replace corrupted message by a sentinel value. It’s a special-purpose record (e.g: null, None, Json.Null, etc ...) Sentinel Value Pattern → G H
54. 54 @loicmdivad @XebiaFr We need to turn the deserialization process into a pure transformation that cannot crash To do so, we will replace corrupted message by a sentinel value. It’s a special-purpose record (e.g: null, None, Json.Null, etc ...) This allows downstream processors to recognize and handle such sentinel values → G H G H Sentinel Value Pattern
55. 55 @loicmdivad @XebiaFr We need to turn the deserialization process into a pure transformation that cannot crash To do so, we will replace corrupted message by a sentinel value. It’s a special-purpose record (e.g: null, None, Json.Null, etc ...) This allows downstream processors to recognize and handle such sentinel values With Kafka Streams this can be achieved by implementing a Deserializer → G H G H null Sentinel Value Pattern
56. 56 @loicmdivad @XebiaFr
57. 57 @loicmdivad @XebiaFr
58. 58 @loicmdivad @XebiaFr @loicmdivad @XebiaFr
59. 59 @loicmdivad @XebiaFr
60. 60 @loicmdivad @XebiaFr @loicmdivad @XebiaFr
61. 61 @loicmdivad @XebiaFr
62. 62 @loicmdivad @XebiaFr 62 @loicmdivad @XebiaFr Sentinel Value Pattern By implementing a custom serde we can create a safe . Downstreams now receive a sentinel value indicating a deserialization error. Errors can then be treated correctly, example: monitoring the number of deserialization errors with a custom metric But we lost a lot of information about the error… let’s see a last method
63. 63 @loicmdivad @XebiaFr 63 @loicmdivad @XebiaFr Sentinel Value Pattern By implementing a custom serde we can create a safe . Downstreams now receive a sentinel value indicating a deserialization error. Errors can then be treated correctly, example: monitoring the number of deserialization errors with a custom metric But we lost a lot of information about the error… let’s see a last method Take Away
64. 64 @loicmdivad @XebiaFr
65. 65 @loicmdivad @XebiaFr Stream Topology Streaming APP
66. 66 @loicmdivad @XebiaFr
67. 67 @loicmdivad @XebiaFr Dead Letter Queue Pattern In this method we will let the deserializer fail. For each failure we will send a message to a topic containing corrupted messages. Each message will have the original content of the input message (for reprocessing) and additional meta data about the failure. With Kafka Streams this can be achieved by implementing a DeserializationExceptionHandler Streaming APP dead letter queue input topic output topic
68. 68 @loicmdivad @XebiaFr
69. 69 @loicmdivad @XebiaFr
70. 70 @loicmdivad @XebiaFr
71. 71 @loicmdivad @XebiaFr
72. 72 @loicmdivad @XebiaFr Fill the headers with some meta data 01061696e0016536f6d6500000005736f6d65206f Value message to hexa
73. 73 @loicmdivad @XebiaFr
74. 74 @loicmdivad @XebiaFr Take Away
75. 75 @loicmdivad @XebiaFr @loicmdivad @XebiaFr
76. 76 @loicmdivad @XebiaFr
77. 77 @loicmdivad @XebiaFr 414554 = AET = Australia/Sydney
78. 78 @loicmdivad @XebiaFr Stream Topology Streaming APP
79. 79 @loicmdivad @XebiaFr 79 @loicmdivad @XebiaFr Dead Letter Queue Pattern You can provide your own implementation of . This lets you use the Producer API to write a corrupted record directly to a quarantine topic. Then you can manually analyse your corrupted records ⚠Warning: This approach have side effects that are invisible to the Kafka Streams runtime.
80. 80 @loicmdivad @XebiaFr 80 @loicmdivad @XebiaFr Dead Letter Queue Pattern You can provide your own implementation of . This lets you use the Producer API to write a corrupted record directly to a quarantine topic. Then you can manually analyse your corrupted records ⚠Warning: This approach have side effects that are invisible to the Kafka Streams runtime. Take Away
81. 81 @loicmdivad @XebiaFr
82. 82 @loicmdivad @XebiaFr 82 @loicmdivad @XebiaFr Links XKE-RATATOUILLE CONFLUENT FAQ
83. 83 @loicmdivad @XebiaFr 83 @loicmdivad @XebiaFr Related Post Kafka Connect Deep Dive – Error Handling and Dead Letter Queues - by Robin Moffatt Building Reliable Reprocessing and Dead Letter Queues with Apache Kafka - by Ning Xia Handling bad messages using Kafka's Streams API - answer by Matthias J. Sax
84. 84 @loicmdivad @XebiaFr 84 @loicmdivad @XebiaFr Conclusion When using Kafka, deserialization is the responsibility of the clients. These internal errors are not easy to catch When it’s possible, use Avro + Schema Registry When it’s not possible, Kafka Streams applies techniques to deal with serde errors: - DLQ: By extending a - Sentinel Value: By extending a
85. 85 @loicmdivad @XebiaFr @loicmdivad @XebiaFr
86. 86 @loicmdivad @XebiaFr 86 @loicmdivad @XebiaFr Images Photo by Michele Blackwell on Unsplash Photo by rawpixel on Unsplash Photo by João Marcelo Martins on Unsplash Photo by Jordane Mathieu on Unsplash Photo by Brooke Lark on Unsplash Photo by Jakub Kapusnak on Unsplash Photo by Melissa Walker Horn on Unsplash Photo by Aneta Pawlik on Unsplash
87. 87 @loicmdivad @XebiaFr 87 @loicmdivad @XebiaFr With special thanks to Robin M. Sylvain L. Giulia B.
88. 88 @loicmdivad @XebiaFr How the generator works?
89. 89 @loicmdivad @XebiaFr Pure HTML Akka Http Server Akka Actor System Kafka Topic Exercise1 Exercise2 Me, clicking everywhere Akka Stream Kafka