ReferenceGlossary

Chaos engineering

Injecting failures on purpose, in a controlled way, to find out how a system really behaves: kill a node, add latency, drop a dependency. It turns a guess about a weak spot into a measured one.

Reliability

In Quorum#

Chaos engineering research adds a Chaos switch to the Team panel. While it is on, the Design’s weakest Node fails 20% of its calls for 10 s every 120 s, and the Company earns 1 Failure Insight a second from the experiment. The switch names the Node the next experiment will hit. The injected `refused` errors show in the Dashboard’s Errors row, beside the real failures.

Design#

Reliability research

Blast radius, Runbook, Metastable failure, Observability

Learn more#