Monitoring and Observability

10 Lektionen in dieser Stufe

1. Metrics, Logs and Traces: What to Collect from a Device
   Choose signals that answer real operational questions. · Lab: List ten questions you need answered about a device and the signal that answers each.
2. Exporting System Metrics with node_exporter
   Publish CPU, memory, disk, temperature and network figures. · Lab: Run the exporter as a service and read its metrics page by hand.
3. Custom Application Metrics in Python
   Instrument your own code with counters, gauges and histograms. · Lab: Add sensor-read duration, error count and queue depth metrics to your service.
4. Scraping and Storing with Prometheus
   Collect metrics from many devices and query them. · Lab: Deploy Prometheus in a container, scrape three devices and write five queries.
5. Dashboards in Grafana
   Design panels that show fleet health at a glance. · Lab: Build a fleet overview and a per-device drill-down with variables.
6. Alert Rules That Do Not Cry Wolf
   Alert on symptoms with sensible thresholds, durations and routing. · Lab: Write alerts for device offline, storage nearly full and sensor stuck, routed to a phone.
7. Shipping Logs Centrally with Loki
   Collect journals from all devices and search them by label. · Lab: Forward journald logs with an agent and find one error across the fleet.
8. Heartbeats and Detecting Silent Devices
   Notice when a device stops reporting, not just when it reports errors. · Lab: Implement a dead-man check that alerts when a heartbeat is missed.
9. Monitoring over Metered and Intermittent Links
   Buffer, batch and downsample to respect data limits. · Lab: Queue metrics during an outage and deliver them after reconnect within a data budget.
10. Defining SLOs and Running a Post-Incident Review
    Set measurable reliability targets and improve from failures. · Lab: Define an availability objective, cause an outage and write a blameless review.

Abschlussprojekt

A monitoring stack with dashboards for system and application health, central logs, alerts that reach your phone and a written service-level objective for your fleet.

Teile für diese Stufe

• 3 × Fleet devices from Level 34
• 1 × Raspberry Pi 5 (8 GB) with NVMe storage as monitoring server
• 1 × M.2 NVMe HAT and SSD from Level 20

In jeder Lektion legst du die passenden Teile direkt in den Warenkorb. Raspberry Pi →