Lightning Talk: What I Wish I’d Known about Fluentd with Kubernetes - Bryan Boreham, Weaveworks
summary
Hands-on notes from tuning Fluentd on a production Kubernetes cluster sending logs to Loggly: install as a daemonset install so logs reach every node, enable buffered output to avoid one-line-at-a-time uploads, add a syslog reader and use Fluentd plugins rather than rewriting a custom logger, and debug by shelling into the container and reading Ruby source. Watch versions (many guides target older 0.12/0.14), and override the image UID to override UID to 0 so it can read root-owned /var/log.
AI-written summary · about
chapters (5)
- 0:00Start
- 0:55Fluentd is awesome
- 2:47Use buffered output
- 3:32It doesn't run as root. Maybe.
- 4:08It's Ruby! Shell in and use the repl
related talks
05mLightning Talk: Flexible Logging Pipelines with Fluentd and Kubernetes
09mLightning Talk: Fluentd at Scale; Keys to successful Logging Syndication
34mFluentd: A Complete Logging Ecosystem for Kubernetes
10mLightning Talk: Fluentd as syslog drain for CloudFoundry
05m⚡Lightning Talk: Learning Kubernetes From Logs: Building a Foundation for Future…
27mUnder the Hood with Fluent Bit Operator: Kubernetes-native Log Processor
09mLightning Talk: Scaling the Fluent Bit Kubernetes Filter in...
06mLightning Talk: Parsing CRI JSON logs with Fluent Bit