How is the role of DevOps changing in a Serverless environment?
summary
Claims serverless doesn't eliminate DevOps but multiplies operational pain: the company migrates to AWS Lambda, uses kappa and apex, runs a Slack bot via ECS + Lambda + SNS + Kinesis and logs to DynamoDB, and ships an open source kappa deploy tool. Problems: developers push from laptops with no deployment records, huge duplicated libraries bloating function zips, rare infinite loops burning money, and no local test harness. Remedies tried: function aliases, queues to avoid throttling, immutable data and traces, and monitoring with cumulative flow diagram.
AI-written summary · about
chapters (11)
- 0:00Start
- 5:22Use Cases
- 8:43Word Generator
- 12:48YeoBot Architecture
- 19:01Immutable Data
- 20:10Queue Depth
- 20:48Tips and Tricks
- 23:14Avoiding Infinte Loops
- 24:16Store your data properly
- 28:11Testing
- 29:02Questions?
related talks
05mUnleash your DevOps with Serverless
37mTHE DEVOPS EVOLUTION
15mThe Ops in Serverless
40mDevOps in the era of serverless computing
23mServerless Ops: What to Do / This is What We Do, When the Server Goes Away
49mMeasuring and increasing developer productivity with the help of Serverless
05m#Serverless + #Devops = BFF
36mTransforming SaaS with Serverless Architectures