3 Ways to Improve Serverless Performance

Join us for an insightful webinar as we explore strategies for optimizing AWS Lambda performance. In this session, Yan Cui, AWS Serverless Hero, and Aviad Mor, Co-founder and CTO of Lumigo, will discuss three approaches that can significantly enhance the efficiency and scalability of your serverless applications.

Discover how to streamline your AWS Lambda functions to achieve better performance, reduce latency, and minimize costs. We’ll share practical tips and best practices gleaned from real-world experiences, empowering you to maximize the potential of serverless computing on AWS.

Transcript

4. Us too!
5. Trace through DynamoDB streams, SNS, SQS, Kinesis, EventBridge, etc.
6. RDS, Redis, Mongodb, Postgres, MS SQL, …
7. RDS, Redis, Mongodb, Postgres, MS SQL, …
9. BEST Serverless Observability Tool!!
10. (one of the biggest serverless environments)
11. Aviad Mor Co-founder and CTO of Lumigo
12. Finding the optimal Lambda settings. 1
13. He who controls the memory, controls the CPU
14. More memory = more CPU = more $ per ms of exec time.
15. 128MB
16. 128MB Don’t blame me for a bit Lambda bill…
17. 128MB Don’t blame me for a bit Lambda bill… I’m just calling an API, 128MB is fi ne!
19. OS paging already kicked in
20. L1 cache L2 cache Main memory SSD disk 0.5 ns 7.0 ns 100.0 ns 10,000,000.0 ns (10 ms)
21. Use 512MB / 1GB memory as default. (even if you don’t need it)
23. Right-size Lambda functions.
24. https://github.com/alexcasalboni/aws-lambda-power-tuning
25. https://github.com/alexcasalboni/aws-lambda-power-tuning
27. Consider return-on-investment.
30. Use billing alarms!
31. Hard to right-size Lambdaliths.
32. Code optimizations. 2
38. Embrace concurrency & parallelism in code
42. Caching. 3
43. Caching is a cheat code for building performant & scalable applications.
44. bit.ly/sls-caching
45. aaronstuyvenberg.com/posts/containers-on-lambda
46. www.usenix.org/system/ fi les/atc23-brooker.pdf
47. bit.ly/social-network-in-4-weeks
48. average 99% cache hit rate
50. Aviad Mor Co-founder and CTO of Lumigo
51. Choosing the right service. 4
52. Cost e ff i ciency at scale.
53. $10.836 1 msg/s for a month, 1KB per msg 1 x 60s x 60m x 24hr x 30days @ $0.014 per mil + 24hrs x 30days @ $0.015 per shard per hr $2.592 SNS SQS EventBridge Kinesis Provisioned 1 x 60s x 60m x 24hr x 30days @ $1.00 per mil 1 x 60s x 60m x 24hr x 30days @ $0.40 per mil 1 x 60s x 60m x 24hr x 30days @ $0.50 per mil $1.037 $1.296 Kinesis On-Demand 1kb x 60s x 60m x 24hr x 30days @ $0.08 per GB ingested + 24hrs x 30days @ $0.04 per stream per hr $28.998
54. $47.088 1,000 msg/s for a month, 1KB per msg 1000 x 60s x 60m x 24hr x 30days @ $0.014 per mil + 24hrs x 30days @ $0.015 per shard per hr $2592.00 SNS SQS EventBridge Kinesis Provisioned 1000 x 60s x 60m x 24hr x 30days @ $1.00 per mil 1000 x 60s x 60m x 24hr x 30days @ $0.40 per mil 1000 x 60s x 60m x 24hr x 30days @ $0.50 per mil $1036.80 $1296.00 Kinesis On-Demand 1000kb x 60s x 60m x 24hr x 30days @ $0.08 per GB ingested + 24hrs x 30days @ $0.04 per stream per hr $226.55
57. Services that charge by up-time is order of magnitude more cost-e ff i cient at scale.
58. Batched vs. Single-events
59. Batched Single-events SNS EventBridge SQS Kinesis
60. Multi-Region vs. Single-Region
61. Aviad Mor Co-founder and CTO of Lumigo
62. Observability. 5
63. Performance optimization is a long game.
64. Observe Orient Decide Act https://en.wikipedia.org/wiki/OODA_loop
65. Observe Orient Decide Act https://en.wikipedia.org/wiki/OODA_loop
66. Aviad Mor Co-founder and CTO of Lumigo
67. 1. Right-sizing Lambda functions. 2. Code optimizations. 3. Caching. 4. Choosing the right service. 5. Observability.
68. Get Started Troubleshooting Your Serverless Databases Today - Free Lumigo.io/signup
69. Questions?