Learn the collocations for profiling performance, identifying bottlenecks, benchmarking results, and setting baselines.
0 / 5 completed
1 / 5
The SRE used py-spy to ___ performance of the Python service and understand where CPU time was being spent.
Profile performance is the standard engineering collocation for using a profiler to collect runtime data on CPU, memory, and call stacks. 'Analyse along' and 'check out' are informal. 'Measure around' is too generic for this specific technique.
2 / 5
After profiling, the team used flame graphs to ___ bottlenecks in the request-handling path.
Identify bottlenecks is the standard performance engineering collocation for locating the specific parts of a system limiting throughput or increasing latency. 'Find along' and 'spot around' are informal. 'Detect out' is redundant.
3 / 5
The team used k6 to run load tests and ___ results against the SLO targets before the release.
Benchmark results is the standard performance testing collocation for measuring and comparing system performance against reference values. 'Compare along' and 'test around' are informal. 'Measure out' is too vague for a formal benchmarking process.
4 / 5
Before the new feature shipped, the team ran load tests to ___ baselines for p50, p95, and p99 latency.
Set baselines is the standard performance and SRE collocation for establishing a reference point against which future performance can be compared. 'Establish along' and 'create around' are informal. 'Record out' does not convey the intentional nature of baseline setting.
5 / 5
The performance engineer recommended a caching layer to ___ load on the database under peak traffic.
Reduce load is the standard performance optimisation collocation for decreasing the demand placed on a resource. 'Lower along' and 'cut around' are informal. 'Decrease out' is not a standard phrase in a performance context.