Serverless in production, an experience report (NDC London 2018)
0 likes618 views
This document provides an experience report on going to production with serverless applications. It discusses several lessons learned including hidden complexities, the need to scale services earlier than with EC2 to avoid latency issues, high costs for unused resources with auto-scaling, and deployment downtimes of up to 30 minutes. It also provides recommendations for serverless deployments including having independent, loosely coupled features that can be deployed separately with zero downtime and no lock-step dependencies.
![if [ "$1" = "deploy" ] && [ $# -eq 4 ]; then
STAGE=$2
REGION=$3
PROFILE=$4
npm install
AWS_PROFILE=$PROFILE 'node_modules/.bin/sls' deploy -s $STAGE -r $REGION
elif [ "$1" = "int-test" ] && [ $# -eq 4 ]; then
STAGE=$2
REGION=$3
PROFILE=$4
npm install
AWS_PROFILE=$PROFILE npm run int-$STAGE
elif [ "$1" = "acceptance-test" ] && [ $# -eq 4 ]; then
STAGE=$2
REGION=$3
PROFILE=$4
npm install
AWS_PROFILE=$PROFILE npm run acceptance-$STAGE
else
usage
exit 1
fi](https://image.slidesharecdn.com/serverlessinprodndclondon-180117185513/85/Serverless-in-production-an-experience-report-NDC-London-2018-98-320.jpg)
Serverless in production, an experience report (NDC London 2018)
- 1. in production an experience reportan experience report what you should know before you go to production ServerlessServerless
- 4. Yan Cui Server Architect Principal Engineer Lead Developer Senior Developer http://theburningmonk.com @theburningmonk Senior Developer
- 5. Yan Cui Server Architect Principal Engineer Lead Developer Senior Developer http://theburningmonk.com @theburningmonk Senior Developer
- 6. Domas Lasauskas AWS user since 2012
- 7. Domas Lasauskas Senior Developer Server Developer Senior Developer Server Developer
- 10. apr, 2016
- 11. hidden complexities and dependencies low utilisation to leave room for traffic spikes EC2 scaling is slow, so scale earlier lots of cost for unused resources up to 30 mins for deployment deployment required downtime
- 12. - Dan North “lead time to someone saying thank you is the only reputation metric that matters.”
- 14. “what would good look like for us?”
- 15. be small be fast have zero downtime have no lock-step DEPLOYMENTS SHOULD...
- 16. FEATURES SHOULD... be deployable independently be loosely-coupled
- 17. WE WANT TO... minimise cost for unused resources minimise ops effort reduce tech mess deliver visible improvements faster
- 18. nov, 2016
- 19. 170 Lambda functions in prod 1.2 GB deployment packages in prod 95% cost saving vs EC2 15x no. of prod releases per month
- 20. time is a good fit
- 21. 1st function in prod! time is a good fit
- 22. ? time is a good fit 1st function in prod!
- 24. Practices ToolsPrinciples what is good? how to make it good? with what?
- 26. 170 functions WOOF! ? ? time is a good fit 1st function in prod!
- 29. rebuilt search
- 30. Legacy Monolith Amazon Kinesis Amazon Lambda Amazon CloudSearch
- 31. Legacy Monolith Amazon Kinesis Amazon Lambda Amazon CloudSearchAmazon API Gateway Amazon Lambda
- 33. Legacy Monolith Amazon Kinesis Amazon Lambda Google BigQuery
- 34. Legacy Monolith Amazon Kinesis Amazon Lambda Google BigQuery 1 developer, 2 days design production (his 1st serverless project)
- 35. Legacy Monolith Amazon Kinesis Amazon Lambda Google BigQuery “nothing ever got done this fast at Skype!” - Chris Twamley
- 36. - Dan North “lead time to someone saying thank you is the only reputation metric that matters.”
- 45. BigQuery
- 46. BigQuery
- 60. TESTING
- 61. amzn.to/29Lxuzu
- 62. Level of Testing 1.Unit do our objects do the right thing? are they easy to work with?
- 64. Level of Testing 1.Unit 2.Integration does our code work against code we can’t change?
- 65. handler
- 66. handler test by invoking the handler
- 67. Level of Testing 1.Unit 2.Integration 3.Acceptance does the whole system work?
- 69. “…We find that tests 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
- 70. “…The second risk is 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
- 71. Don’t Mock Types You Can’t Change Services
- 72. Paul Johnston The serverless approach to testing is different and may actually be easier. http://bit.ly/2t5viwK
- 74. LambdaAPI Gateway DynamoDB Unit Tests
- 75. LambdaAPI Gateway DynamoDB Unit Tests Mock/Stub
- 76. is our request correct? is the request mapping set up correctly?is the API resources configured correctly? are we assuming the correct schema? LambdaAPI Gateway DynamoDB is Lambda proxy configured correctly? is IAM policy set up correctly? is the table created? what unit tests will not tell you…
- 78. most Lambda functions are simple have single purpose, the risk of shipping broken software has largely shifted to how they integrate with external services observation
- 80. But it slows down my feedback loop… IT’S NOT ABOUT YOU!
- 81. …if a service can’t provide you with a relatively easy way to test the interface in reality, then you should consider using another one. Paul Johnston
- 82. “…Wherever possible, an acceptance 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
- 83. Legacy Monolith Amazon Kinesis Amazon Lambda Amazon CloudSearchAmazon API Gateway Amazon Lambda
- 84. Legacy Monolith Amazon Kinesis Amazon Lambda Amazon CloudSearchAmazon API Gateway Amazon Lambda Test Input
- 85. Legacy Monolith Amazon Kinesis Amazon Lambda Amazon CloudSearchAmazon API Gateway Amazon Lambda Test Input Validate
- 86. integration tests exercise system’s Integration with its external dependencies my code
- 87. acceptance tests exercise system End-to-End from the outside my code
- 88. integration tests differ from acceptance tests only in HOW the Lambda functions are invoked observation
- 92. CI + CD PIPELINE
- 93. “the earlier you consider CI + CD, the more time you save in the long run” - Yan
- 94. “…We prefer to have 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
- 95. “deployment scripts that only live on the CI box is a disaster waiting to happen” - Yan
- 96. Jenkins build config deploys and tests unit + integration tests deploy acceptance tests
- 97. build.sh allows repeatable builds on both local & CI
- 98. if [ "$1" = "deploy" ] && [ $# -eq 4 ]; then STAGE=$2 REGION=$3 PROFILE=$4 npm install AWS_PROFILE=$PROFILE 'node_modules/.bin/sls' deploy -s $STAGE -r $REGION elif [ "$1" = "int-test" ] && [ $# -eq 4 ]; then STAGE=$2 REGION=$3 PROFILE=$4 npm install AWS_PROFILE=$PROFILE npm run int-$STAGE elif [ "$1" = "acceptance-test" ] && [ $# -eq 4 ]; then STAGE=$2 REGION=$3 PROFILE=$4 npm install AWS_PROFILE=$PROFILE npm run acceptance-$STAGE else usage exit 1 fi
- 99. Jenkinsfile
- 101. Auto Auto Manual
- 102. LOGGING
- 104. 2016-07-12T12:24:37.571Z 994f18f9-482b-11e6-8668-53e4eab441ae GOT is off air, what do I do now?
- 105. 2016-07-12T12:24:37.571Z 994f18f9-482b-11e6-8668-53e4eab441ae GOT is off air, what do I do now? UTC Timestamp API Gateway Request Id your log message
- 107. Yan Logs are not easily searchable in CloudWatch Logs.
- 108. LOG OVERLOAD
- 109. CENTRALISE LOGS
- 110. CENTRALISE LOGS MAKE THEM EASILY SEARCHABLE
- 111. + + the elk stack
- 112. CloudWatch Logs
- 113. CloudWatch Logs AWS Lambda ELK stack
- 114. CloudWatch Events
- 117. DISTRIBUTED TRACING
- 119. “my followers didn’t receive my new post!” - a user
- 120. where could the problem be?
- 121. correlation IDs* * eg. request-id, user-id, yubl-id, etc.
- 123. kinesis client http client sns client
- 126. Amazon X-Ray
- 127. Amazon X-Ray
- 128. traces do not span over API Gateway
- 131. “where do I install monitoring agents?”
- 132. you can’t
- 133. • invocation Count • error Count • latency • throttling • granular to the minute • support custom metrics
- 134. • same metrics as CW • better dashboard • support custom metrics https://www.datadoghq.com/blog/monitoring-lambda-functions-datadog/
- 138. my code
- 139. my code
- 140. my code internet internet press button something happens
- 141. “how do I batch up and send logs in the background?”
- 142. you can’t (kinda)
- 143. console.log(“hydrating yubls from db…”); console.log(“fetching user info from user-api”); console.log(“MONITORING|1489795335|27.4|latency|user-api-latency”); console.log(“MONITORING|1489795335|8|count|yubls-served”); timestamp metric value metric type metric namemetrics logs
- 144. CloudWatch Logs AWS Lambda ELK stack logs metrics CloudWatch
- 146. DASHBOARDS
- 149. Not Only CloudWatch
- 151. “you really don't want your monitoring system to fail at the same time as the system it monitors” - me
- 152. CONFIG MANAGEMENT
- 153. easily and quickly propagate config changes
- 155. Yan Environment variables make it hard to share configurations across functions.
- 156. Yan Environment variables make it hard to implement fine-grained access to sensitive info.
- 162. SSM Parameter Store
- 163. sensitive data should be encrypted in-flight, and at rest (credentials, connection string, etc.)
- 164. role-based access
- 165. SSM Parameter Store HTTPS role-based access encrypted in-flight
- 166. SSM Parameter Store encrypt role-based access
- 167. SSM Parameter Store encrypted at-rest
- 168. HTTPS role-based access SSM Parameter Store encrypted in-flight
- 170. fetch & cache at Cold Start
- 171. invalidate at interval + signal
- 173. PRO TIPS
- 174. max 75 GB total deployment package size* * limit is per AWS region
- 175. Janitor Monkey
- 178. install Serverless framework as dev dependency at project level dev dependencies are excluded since 1.16.0
- 182. Amazon X-Ray 1st invocation 2nd invocation cold start
- 189. what about type safety?
- 191. complexity ceiling of a Node.js app complexity
- 192. complexity ceiling of a Node.js app complexity referential transparency immutability as default type inference option types union types …
- 193. for managing complexity complexity ceiling of a Node.js app complexity referential transparency immutability as default type inference option types union types …
- 194. complexity ceiling of a Node.js app complexity complexity ceiling of a Node.js Lambda function
- 195. if you can limit the complexity of your solution, maybe you won’t need the tools for managing that complexity. me
- 196. AVOID COLDSTARTS
- 197. CloudWatch Event AWS Lambda
- 198. CloudWatch Event AWS Lambda ping ping ping ping
- 199. CloudWatch Event AWS Lambda ping ping ping ping
- 200. CloudWatch Event AWS Lambda ping ping ping ping HEALTH CHECKS?
- 203. max 5 mins execution time
- 204. USE RECURSION FOR LONG RUNNING TASKS
- 206. “AWS Lambda polls your stream and invokes your Lambda function. Therefore, if a Lambda function fails, AWS Lambda attempts to process the erring batch of records until the time the data expires…” http://docs.aws.amazon.com/lambda/latest/dg/retries-on-errors.html
- 207. should function fail on partial/any failures?
- 208. SNS Kinesis SQS after 3 attempts share processing logic events are processed in chronological order failed events are retried out of sequence
- 212. “Each shard can support up to 5 transactions per second for reads, up to a maximum total data read rate of 2 MB per second.” http://docs.aws.amazon.com/streams/latest/dev/service-sizes-and-limits.html
- 213. “If your stream has 100 active shards, there will be 100 Lambda functions running concurrently. Then, each Lambda function processes events on a shard in the order that they arrive.” http://docs.aws.amazon.com/lambda/latest/dg/concurrent-executions.html
- 214. when no. of processors goes up…
- 215. ReadProvisionedThroughputExceeded can have too many Kinesis read operations…
- 216. ReadRecords.IteratorAge unpredictable spikes in read ‘latency’…
- 219. clever, but costly
- 220. for subsystems that don’t have to be realtime, or are task- based (ie. order doesn’t matter), consider other triggers such as S3 or SNS.me
- 222. API Gateway and Kinesis Authentication & authorisation (IAM, Cognito) Testing Running & Debugging functions locally Log aggregation Monitoring & Alerting X-Ray Correlation IDs CI/CD Performance and Cost optimisation Error Handling Configuration management VPC Security Leading practices (API Gateway, Kinesis, Lambda) Step Functions Serverless design patterns https://bit.ly/aws-lambda-in-motion
- 223. API Gateway and Kinesis Authentication & authorisation (IAM, Cognito) Testing Running & Debugging functions locally Log aggregation Monitoring & Alerting X-Ray Correlation IDs CI/CD Performance and Cost optimisation Error Handling Configuration management VPC Security Leading practices (API Gateway, Kinesis, Lambda) Step Functions Serverless design patterns get 40% off with: ytcui https://bit.ly/aws-lambda-in-motion