Advanced Observability: Real-World Monitoring and Logging

Explore observability in distributed environments, focusing on monitoring, logging, and tracing to improve reliability. Learn how site reliability engineers use Prometheus and Kubernetes to detect, alert, and diagnose issues.

Explore Prometheus architecture, where exporters collect metrics from machines or containers, store labeled time-series data in a database, and feed Grafana dashboards, PromQL queries, and Alertmanager for monitoring and alerting.

Set up Prometheus on two AWS machines—one to run the server, one to monitor—scraping metrics every 15 seconds, and view the built-in dashboard across clouds.

Monitor an external machine with Prometheus by installing node exporter, exposing metrics at 9100, and configuring Prometheus to scrape them into the time series database.

Prometheus is widely used and the de facto monitoring tool, especially with Kubernetes, with Windows exporters and MSI installers, enabling custom exporters and 15-second scraping intervals.

Learn to deploy Prometheus and node exporter on two AWS machines, scrape metrics from port 9100 every 15 seconds, and monitor a web server.

Install nginx and the nginx prometheus exporter to expose nginx metrics via a dedicated http endpoint for Prometheus to scrape. Enable status or substatus and feed data to Prometheus.

Learn how Prometheus assigns default labels, such as job and instance, to metrics scraped from multiple targets (like node exporter), enabling attribution of data to each machine.

Visualize Prometheus metrics from node exporter and nginx exporter in a time series database. Use PromQL queries and Grafana dashboards to transform data into meaningful insights.

Install Grafana on an AWS Ubuntu instance, start the service, and access it through port 3000; connect to Prometheus and build dashboards.

Create dashboards in Grafana from Prometheus data, building multiple panels to monitor CPU usage, memory, network, and processes, using queries, rate, and sum for insights.

Learn how to add a memory usage panel in Grafana, query memory metrics with PromQL, and interpret free memory, cache, and buffer in a dashboard for a web server.

Explore Grafana's dashboard library to import pre-built node exporter dashboards, visualize Prometheus-collected CPU, memory, and network metrics, and quickly set up web server monitoring.

Learn to manage Grafana dashboards as code by exporting and provisioning JSON files, using YAML provisioning, and maintaining dashboards from the command line rather than the user interface.

Generate comprehensive monitoring data by deploying exporters (node, nginx, mysql) or using SDKs and custom scripts, enabling Prometheus to scrape metrics via HTTP endpoints for Grafana dashboards.

Discover how Prometheus uses pushgateway to collect metrics from short-lived targets by pushing data to a gateway, configuring targets, labels, and scraping via Prometheus and Grafana.

Learn to build dynamic Grafana dashboards by importing a Node Exporter dashboard, configuring multiple Prometheus data sources, and using variables like host and instance to filter by machine.

Learn how to run Prometheus inside a Kubernetes cluster with exporters, connect to Grafana, manage configurations with config maps and secrets, and plan for multi Prometheus setups.

Learn to deploy a Prometheus observability stack on Kubernetes with node exporter, an API server exporter, Prometheus stateful set, Grafana, pushgateway, alertmanager, and Helm stack.

Install the Prometheus stack on a Kubernetes cluster using Helm. Customize Grafana and Prometheus via a values file, enable public access, and set admin password.

Configure and manage the Prometheus stack in Kubernetes using helm, understanding deployments, daemonsets, and statefulsets, and expose Grafana via port forwarding while editing config maps for Grafana and dashboards.

Use kubectl port-forward to expose grafana and prometheus from your kubernetes cluster to your laptop, then access grafana at localhost:8080 and prometheus at localhost:9090.

Explore default monitoring configurations in a Kubernetes stack, where Prometheus auto-discovers targets via service monitors and Grafana offers prebuilt dashboards for api server, node exporter, and cluster components.

Configure a kubernetes app with the Prometheus nginx exporter as a sidecar, expose metrics at /metrics on port 9113, and enable Prometheus scraping via a service monitor.

Explore Prometheus rules to record custom metrics in the Prometheus time series database using PromQL, and learn how to configure rules in Kubernetes, YAML, and Helm environments.

Learn how to balance availability and latency using SLI, SLO, and SLA concepts, measure with Prometheus, and implement alerting in Kubernetes and nginx-based services.

Automate Prometheus rule generation for SLO based monitoring with Sloth, using nginx metrics and multi-window error rates, with alerting and end-of-month availability goals.

Apply and visualize SLO based rules using Prometheus recording and alerting rules, with PromQL queries for error rates across time windows, track error budget and burn rate on Grafana dashboards.

Prioritize services by SLAs and availability, visualize service level objectives with Grafana for Prometheus targets, and decide releases based on burn rate and customer impact.

Explore Grafana Loki for centralized logging, detailing log collection with promtail, gateway coordination, and central storage, then visualize logs in Grafana dashboards using logQL and labels.