Downloaded 91 times
Uploaded byYan Cui
AWS Lambda from the Trenches
The document discusses best practices and insights for deploying AWS Lambda in production, emphasizing the importance of testing, monitoring, and configuration management. It highlights strategies for cost efficiency, error handling, and serverless architecture design, while also stressing the need for centralizing logs and managing serverless functions effectively. Key aspects include leveraging end-to-end testing, maintaining clear documentation, and understanding the limits and capabilities of AWS Lambda services.
More Related Content
AWS Lambda from the Trenches
- 1. + = AWS LAMBDAFROM THE TRENCHES what you should know before you go to production
- 2. hi, my nameis Yan Cui
- 3.
- 5. - Dan North “leadtime to someone saying thank you is the only reputation metric that matters.”
- 7. security complexity OUTSIDE the code deployment loadbalancing caching monitoring config management https://www.infoq.com/presentations/complexity-simplicity-esb centralised logging elastic scaling setup server
- 10. THERE IS NOSERVER
- 11.
- 12.
- 13.
- 14. - Dan North “leadtime to someone saying thank you is the only reputation metric that matters.”
- 15.
- 16. not paying for idleservers
- 17.
- 18.
- 20.
- 21.
- 22. Legacy Monolith AmazonKinesis Amazon Lambda
- 23. Legacy Monolith AmazonKinesis Amazon Lambda Amazon CloudSearchAmazon API Gateway Amazon Lambda
- 24.
- 25. Legacy Monolith AmazonKinesis Amazon Lambda Google BigQuery
- 26. Legacy Monolith AmazonKinesis Amazon Lambda Google BigQuery 1 developer, 2 days design production (his 1st serverless project)
- 27. Legacy Monolith AmazonKinesis Amazon Lambda Google BigQuery “nothing ever got done this fast at Skype!” - Chris Twamley
- 28. - Dan North “leadtime to someone saying thank you is the only reputation metric that matters.”
- 29.
- 30. Amazon Lambda GrapheneDBAmazonAPI Gateway Amazon API Gateway Amazon Lambda Facebook Graph API
- 31.
- 32.
- 33.
- 34.
- 35.
- 36.
- 37.
- 39. Amazon Lambda Amazon Kinesis AmazonIOT Amazon IOT
- 40. “I thought ofobjects being like biological cells and/or individual computers on a network, only able to communicate with messages.” - Alan Kay
- 41. Amazon Lambda Amazon Kinesis AmazonIOT Amazon IOT
- 42. “OOP to memeans only messaging, local retention and protection and hiding of state- process, and extreme late- binding of all things.” - Alan Kay
- 43.
- 44. Level of Testing 1.Unit doour objects do the right thing? are they easy to work with?
- 46. Level of Testing 1.Unit 2.Integration doesour code work against code we can’t change?
- 47.
- 48. handler test by invoking thehandler
- 49. Level of Testing 1.Unit 2.Integration 3.Acceptance doesthe whole system work?
- 50.
- 51. Level of Testing unit integration acceptance cando all 3 with Lambda
- 52. “…We find thattests that mock external libraries often need to be complex to get the code into the right state for the functionality we need to exercise. The mess in such tests is telling us that the design isn’t right but, instead of fixing the problem by improving the code, we have to carry the extra complexity in both code and test…” Don’t Mock Types You Can’t Change
- 53. “…The second riskis that we have to be sure that the behaviour we stub or mock matches what the external library will actually do… Even if we get it right once, we have to make sure that the tests remain valid when we upgrade the libraries…” Don’t Mock Types You Can’t Change
- 54. Don’t Mock TypesYou Can’t Change Services
- 55. “…Wherever possible, anacceptance test should exercise the system end-to- end without directly calling its internal code. An end-to-end test interacts with the system only from the outside: through its interface…” Testing End-to-End
- 56. Legacy Monolith AmazonKinesis Amazon Lambda Amazon CloudSearchAmazon API Gateway Amazon Lambda
- 57. Legacy Monolith AmazonKinesis Amazon Lambda Amazon CloudSearchAmazon API Gateway Amazon Lambda Test Input
- 58. Legacy Monolith AmazonKinesis Amazon Lambda Amazon CloudSearchAmazon API Gateway Amazon Lambda Test Input Validate
- 59. “…We prefer tohave the end-to-end tests exercise both the system and the process by which it’s built and deployed… This sounds like a lot of effort (it is), but has to be done anyway repeatedly during the software’s lifetime…” Testing End-to-End
- 60. Jenkins build configdeploys and tests unit + integration tests deploy acceptance tests
- 61. build.sh allows repeatablebuilds on both local & CI
- 62.
- 63.
- 64.
- 66. PRO TIP don’t ignore_meta
- 67.
- 68.
- 72.
- 73. GOALS sensitive data are encryptedat rest (credentials, connection string, etc.)
- 74. GOALS has to workon CI
- 75.
- 76. hand-rolled with KMS (encryptedat rest)
- 77.
- 78.
- 79.
- 80.
- 82.
- 83.
- 84. set goals choose away document
- 85.
- 86. set goals choose away evaluate document
- 87. set goals choose away evaluate document
- 88. set goals choose away evaluate document share
- 89.
- 91. 2016-07-12T12:24:37.571Z 994f18f9-482b-11e6-8668-53e4eab441ae GOT isoff air, what do I do now?
- 92. 2016-07-12T12:24:37.571Z 994f18f9-482b-11e6-8668-53e4eab441ae GOT isoff air, what do I do now? UTC Timestamp API Gateway Request Id your log message
- 93. organised by Function+ Version
- 94.
- 95.
- 96. CloudWatch Logs AWSLambda LogStash ElasticSearch
- 97. CloudWatch Logs AWSLambda LogStash ElasticSearch AWS Elasticsearch
- 98. CloudWatch Logs AWSLambda LogStash ElasticSearch AWS Elasticsearch Elastic Cloud
- 99. CloudWatch Logs AWSLambda LogStash ElasticSearch AWS Elasticsearch Elastic Cloud ?
- 100.
- 101.
- 102. PRO TIP set updashboards
- 103. PRO TIP don’t forgetto set up alarms
- 104.
- 105.
- 106. “how do Ireturn HTTP error codes?”
- 109. { “status” :404, “errorMessage” : ”oops” }
- 110. { “status” :404, “errorMessage” : ”oops” }
- 111. s-templates.json { “status” :404, “errorMessage” : ”oops” }
- 112. PRO TIP map timeoutsto 504
- 113. every Lambda functionhas a timeout setting
- 114. use error regexto map it to a HTTP 504
- 115.
- 116. PRO TIP avoid using128mb setting for production
- 117.
- 118. PRO TIP proactively timeout your function
- 119. “what’s the retrystrategy with Kinesis and SNS?”
- 120. “…If the invocationfor one record times out, is throttled, or encounters any other error, Lambda will retry until it succeeds (or the record reaches its 24-hour expiration) before moving on to the next record…” http://aws.amazon.com/lambda/faqs
- 121. • do nothing •swallow errors • track retry count effort
- 122. • retry forever •no retry • retry N times
- 123. PRO TIP use localstate to track no. of retries; move on after N retries
- 124. PRO TIP record CloudWatch metricsfor error count; alarm if necessary
- 125.
- 126.
- 127. functions are unloadedif idle for a while
- 128. noticeable cold starttime (package size matters)
- 129. CloudWatch Event AWSLambda
- 130. CloudWatch Event AWSLambda ping ping ping ping
- 131. CloudWatch Event AWSLambda ping ping ping ping
- 132. CloudWatch Event AWSLambda ping ping ping ping HEALTH CHECKS?
- 133.
- 134. functions are recycled everyfew hours
- 135. functions are recycled everyfew hours
- 136. PRO TIP don’t makehard assumptions about function lifetime
- 137.
- 138. max 50 MBdeployment package size
- 139. max 50 MBdeployment package size max 75 GB total deployment package size* * limit is per AWS region
- 140.
- 141.
- 142. max 5 minsexecution time
- 144. max 6 MBrequest payload size* max 6 MB response payload size * for a request-response event type
- 145. default max 100concurrent executions* * soft-limit, can be raised via support ticket
- 146.
- 148.
- 149.
- 150.
- 151. IBM OpenWhisk Amazon LambdaAzure Web Functions Google Cloud Functions competition faster innovation lower prices
- 153.