Monitoring and metrology for containers
summary
Presents a concrete monitoring architecture for containers: use Docker log drivers to send logs via syslog into Kafka and index in Elasticsearch, and collect host metrics with Telegraf/collectd into Prometheus or InfluxDB visualized with Grafana. Recommends avoiding in-container agents by tagging context at ingestion, discusses retention/capacity tradeoffs and Kafka for high-throughput use (millions of messages/sec), and gives actionable configuration tips and operational pitfalls for deployable patterns.
AI-written summary · about
chapters (12)
- 0:00Container monitoring intro
- 5:58Containers in architectures
- 11:35Monitoring challenges
- 17:11New monitoring paradigm
- 22:27Centralizing logs with syslog
- 26:17Docker logs and journalctl
- 33:42System metrics collection
- 36:26Business metrics and dev
- 41:00Metrics tools selection
- 43:26Proactive alerts and trends
- 47:16Tool simplification and automation
- 49:04Data volume and retention
related talks
46mMonitoring and metrology for containers
23mHow to Monitor Application Performance in a Container-Based World
29mKubernetes-Defined Monitoring [I]
33mObservability-First Kafka: Engineering Visibility at Scale
42mUnified Monitoring of Containers and Microservices [I]
33mThe Patterns of Distributed Logging and Containers
37mYou’re Monitoring Kubernetes Wrong
01h 00mMonitor your containers with the Elastic Stack