Engineering for Reliability
Observing container environments with Cloud Operations
Twenty-three substantive episodes on SLOs, telemetry, alerting, investigation, and operations
Twenty-three 2021–2022 episodes connect user-facing reliability targets to metrics, logs, traces, and alerts. Examples show instrumentation, Prometheus collection, quota monitoring, GKE investigation, and observability cost controls.
In this collection Browse 23 summaries 10 of 23
Which logs, measurements, and request traces can you use to investigate a containerized application? This 2021 comparison shows the data available in Google Kubernetes Engine (GKE), Cloud Run, and the then-named Anthos platform, including where application code needs extra instrumentation 00:00–07:22(opens in a new tab).
Key Points Covered
- The operating model determines what you manage: GKE exposes Kubernetes cluster controls; Cloud Run manages infrastructure and scales in response to traffic; Anthos is presented as a way to manage applications across cloud and other environments 00:00–02:05(opens in a new tab).
- GKE exposes infrastructure and workload evidence: With the demonstrated settings, system and workload logs plus cluster and container metrics are collected, and dashboards connect alerts to pods and containers 02:05–04:10(opens in a new tab).
- Logs and metrics do not automatically supply request traces: The presenter says the GKE workloads shown need OpenTelemetry instrumentation for tracing. By contrast, the Cloud Run example opens a trace showing an individual request's latency and dependencies, alongside automatically collected request counts and latency measurements 03:08–06:18(opens in a new tab).
- Hybrid clusters need infrastructure and application views: The Anthos example checks cluster-management components, node resource use, and workload measurements, then filters container logs to one application's pod 05:12–07:22(opens in a new tab).
Full video: https://www.youtube.com/watch?v=ilY4-Aw1Z2M(opens in a new tab)