2020-03-28, 14:05–14:35, Room 2
A serverless framework provides clear benefits for developers such as easier deployments and scalability. On the other hand, knowing what is really going on is difficult due to the additional abstraction layer and traditional observability methods might not fully work.
While platform metrics, like response time and number of requests, are managed by the serverless framework, collecting application metrics reliably is still a challenge when workloads scale up and scale down frequently.
Tracing and logging also need to be taken into account and tuned accordingly to provide a complete overview when your requests are being processed by multiple functions and debugging capabilities are limited.
A live demo using OpenFaaS, Prometheus, Loki, Jaeger and Grafana to monitor an ecommerce application will show the capabilities of a setup that will improve the observability of the whole platform and is able to scale.
In this talk I will present how you can get insights from your serverless platform and the importance of having great observability.