Observability engineering has its own precise collocations. Practise the phrases SREs and developers use when building monitoring systems.
0 / 5 completed
1 / 5
Fill in: 'The SRE team spent a week ___ the service to expose latency at every function boundary.'
We 'instrument code' — 'instrument' is the observability-specific collocation for adding telemetry (logs, metrics, traces) to source code. 'Monitor' describes an ongoing operational activity; 'measure' focuses on quantity; 'track' is informal.
2 / 5
Fill in: 'Each microservice should ___ custom metrics to Prometheus at least every 15 seconds.'
We 'emit metrics' — 'emit' is the standard observability collocation for a service generating and outputting metric data. 'Send metrics' and 'push metrics' are also acceptable; 'export metrics' is used specifically for Prometheus exporter patterns but 'emit' is the broader idiomatic term.
3 / 5
Fill in: 'After the incident, the team ___ new dashboards to visualise error rates per endpoint.'
We 'set up dashboards' — 'set up' is the standard collocation for configuring and deploying monitoring dashboards. 'Created dashboards' is grammatically fine but less idiomatic in DevOps/SRE speech; 'built' and 'made' are informal.
4 / 5
Fill in: 'The observability platform lets you ___ traces across services to pinpoint a slow database call.'
We 'correlate traces' — 'correlate' is the distributed-tracing collocation for connecting spans across service boundaries using trace IDs. 'Connect traces' and 'link traces' are acceptable but less precise; 'match traces' implies a search operation rather than causal correlation.
5 / 5
Fill in: 'Configure Alertmanager to ___ on latency thresholds exceeding 500 ms for the payment service.'
We 'alert on thresholds' — 'alert on' is the standard observability collocation for defining conditions that fire notifications. 'Notify on thresholds' is close but 'notify' focuses on the delivery; 'trigger on' is informal; 'warn on' is too mild for production alert language.