NGINX meetup talk, Feb 2016

Transcript

1. NGINX Metrics + Monitoring NGINX Meetup, SF February 25, 2016
2. $ whoami John Matson Technical Content @datadoghq @jmtsn Me in sunny San Francisco* *(Nayarit, MX)
3. Agenda • Motivation • What we care about • What you get from /nginx_status (NGINX OSS) • What you get from NGINX Plus • What you get from logs • How Datadog can help
4. Collecting data is cheap, but not having it when you need it is expensive. Motivation
5. Focus on Work metrics Is NGINX servicing requests? How fast? Are there errors?
6. /nginx_status (OSS) • Key work metrics: 
 requests (throughput), dropped connections (error) • Pros: Easy to set up. Free. • Cons: Incomplete coverage. No latency or HTTP codes.
7. NGINX Plus • Key work metrics: 
 requests (throughput), dropped connections (error), 4xx/5xx responses (error) • Pros: Lots of metrics. JSON! • Cons: No latency metrics. demo.nginx.com/status.html
 demo.nginx.com/status
8. NGINX logs • Key work metrics: 
 4xx/5xx responses (error), latency (performance) • Pros: One-line config change. Customizable format.
 HTTP/1 vs HTTP/2 requests! • Cons: Low signal/noise ratio. 
 Need to parse logs. Request Response code Response time
9. Recap: Metric scorecard
10. Where Datadog fits in • Capture/graph/alert on any NGINX or NGINX Plus metric • Can parse NGINX logs too (but not out of the box) • Correlate NGINX metrics with system-level metrics (load, CPU, etc.) or with app metrics • Alert to PagerDuty, Slack, etc. NGINX dashboard Partial list of 120+ supported integrations
11. Further reading • How to Monitor NGINX
 (three-part series):
 dtdg.co/monitor-nginx • Monitoring 101
 (work/resource metrics, etc):
 dtdg.co/monitoring-101-data
12. Thank you! Questions now or later? @jmtsn john@datadoghq.com
13. Supplementary material
16. NGINX connections
17. NGINX Plus connections