One idea at a time
Field guides
Ten illustrated, two-minute guides. Learn a model, inspect an example, then try a lab.
-
Reading service signals
A faster response can hide a worse outcome. Read demand, errors and latency together.
-
Every attempt is load
Count downstream work, not just users. See how retries can fill a queue.
-
Networking on the bridge
A name can resolve without a working destination. Separate DNS, routing and readiness.
-
Resource pressure
Find the limit that actually applies. Host headroom cannot override a smaller boundary.
-
Safe rollouts and recovery
Compare canary and control. A green deployment does not prove operations recovered.
-
Incident investigation
Make a prediction, test one change, and record what the result really shows.
-
Coordination and exactly-once
Each step reports its own success. Trace one effect.
-
Least privilege that works
Read the denial. Grant only what is needed.
-
Automation that converges
A second run should change nothing.
-
Searching lab logs
Ask your logs a question.
No guide matches that search
Try a shorter term or clear the filters.
Concept notes
-
Packets, state and translation
Link MTU, stateful firewalls, source and destination NAT, policy routing.
-
Pod traffic, names and shutdown
Egress policy, Service spreading, StatefulSet names, graceful termination.
-
Agreement and acknowledgement
Quorum, leases, store quotas, synchronous commits.
-
Measuring load and health
Load models, SLI ratios, profiling.
-
Proxies and redirects
Accepted configuration, redirect methods.
Ready to practise? Choose a lab. Briefings open before any session starts.