Life of a startup...
summary
Startup scales a Spring Boot monolith into a Mesos/Marathon Docker cluster with Consul service discovery, Kafka-backed ingestion and Cassandra for large visitor profiles, plus Graylog for alerts and Prometheus + Grafana for metrics. Elasticsearch for analytics proves too heavy, so ClickHouse replaces it and shrinks three months of event storage from ~3 TB to ~100 GB while dropping r95 latency from 4.5s to 300ms. Centralized logging, health checks and automated restarts make the platform resilient; initial orchestration bootstrap costs about 1 FTE for 3 months.
AI-written summary · about
chapters (15)
- 0:00Introduction
- 1:13How does it work
- 2:14How do we start
- 3:48Challenges
- 6:46Solution
- 8:03Peak traffic
- 9:48Elastic Search
- 11:56Mesos
- 15:27Service Discovery
- 18:03Monitoring
- 20:39Cluster Park
- 22:28Monitoring Logging
- 26:43Elasticsearch
- 30:50Conclusion
- 32:57Questions
related talks
51mReal world microservices: Starting-up from scratch!
51mMicroservices and Microfrontends: Scaling, Learning, Growing
28mMicroliths: how to avoid traps in your stack and your culture
46mBuilding on spring boot lastminute.com microservices way
25mMonitoring a JHipster microservice architecture in practice
36mHow to run like a startup with enterprise Kubernetes
48mContainer: check! Aber wohin mit Persistent Data, Big Data oder Fast Data?!
48mTransforming a monolith to an API platform