Carl Bergquist (Grafana) and Neil Gehani (Weaveworks) discuss best practices on how to get started with monitoring your application. Start capturing metrics that matter, aggregate and visualize them in a useful way that allows for identifying bottlenecks and preventing incidents before they happen.
Navigating the Deluge_ Dubai Floods and the Resilience of Dubai International...
Why observability matters - now and in the future (w/guest Grafana)
1. Why Observability Matters?
now and in the future
Monitoring containerized services with Weave Cloud and Grafana
Neil Gehani - Director of Product, Weaveworks - @GehaniNeil
1
2. ● Who is this for? - Software Development/DevOps teams building modern software in containers
● Why Monitoring?
● Weave Cloud Monitor - Applications/Services Monitoring (powered by Prometheus)
● Weave Cloud Ops - Infrastructure Monitoring (powered by Prometheus and Grafana)
● Grafana - Visualizing Prometheus Metrics
What the Webinar is about
2
5. How will dev teams find and fix problems in this system?
5
Dev Team 1 Dev Team 2
Dev Team 3 Dev Team 4
DevOps Team 5 DevOps Team 6
DevOps Team 7 Dev Team 8
7. ●Deploy
○Dry run deploys - try before deploy
○Pull based Ops (GitOps)
○Fully automated or manual including rollbacks
●Explore
○Visual Map and Debug Applications -> Clusters, Pods, Containers, Services
○Time Travel - Compare states at different points in time
●Monitor
○Real-time and historical service, node, cluster monitoring
○Notebooks - Create, Modify, Save Prometheus queries (PromQL)
○Secure, Scalable, Multi-Tenant - as a Service
○Time Travel - Go back in time to see how applications behaved then and compare to now
Weave Cloud - Deploy . Explore . Monitor
7
8. ● Monitor resource consumption
● Observe scaling & performance characteristics
● Alert and Notify when something goes wrong
● Explore and Debug problems
Why Monitor? - Understand Application Behavior
8
Deliver Reliable
Value
Continuous
Improvement
9. Deploy . Observe . Operate
9
01 Automated or Manual with rollbacks.
Deployment event metrics correlated
Deploy
02 Explore, Monitor, Alert, Notify, Log
Observe
03 Manage apps resource usage on clusters.
Add capacity on-demand to scale
automatically
Operate
15. ● Scalable and simple solution for open source Prometheus
● Default monitoring solution as part of CNCF - trusted community support
● K8s cluster metrics available OOB
● Easy to instrument your app - supports all major client libraries
● Easy to customize displaying metrics using PromQL
● Easy to alert based on metrics OOB
● Many visualizations options available - e.g. Grafana
Why Weave Cloud Monitoring for Dev/DevOps Teams?
15