Build fluency in the vocabulary of one tenant's workload degrading performance for others on shared infrastructure.
0 / 5 completed
1 / 5
At standup, a dev mentions that one tenant's batch job on shared infrastructure is consuming so much disk I/O that every other tenant sharing the same physical hardware is experiencing degraded performance, even though their own workloads haven't changed. What is this problem called?
A noisy neighbor is exactly this: it describes a tenant or workload on shared infrastructure that consumes so much of a shared resource, like disk I/O, CPU, or network bandwidth, that it degrades performance for other tenants sharing the same underlying hardware, even though those other tenants' own workloads haven't changed. A hash collision is an unrelated hash-table concept about two keys sharing a bucket. This one-tenant-degrades-everyone-else's-performance pattern is exactly why noisy neighbor problems are a central concern in any multi-tenant, shared-infrastructure environment.
2 / 5
During a design review, the team adds per-tenant resource quotas and I/O throttling to a shared platform, specifically because capping how much disk I/O any single tenant can consume prevents that tenant's batch jobs from degrading performance for every other tenant on the same hardware. Which capability does this provide?
Per-tenant resource quotas here provide resource isolation between tenants on shared infrastructure, since capping how much disk I/O any single tenant can consume prevents one tenant's workload from degrading performance for every other tenant sharing the same hardware. Leaving resource consumption uncapped for every tenant means any one tenant's heavy workload can freely degrade performance for everyone else on that hardware. This cap-and-isolate-resource-consumption behavior is exactly why per-tenant quotas and throttling are the standard fix for noisy neighbor problems.
3 / 5
In a code review, a dev notices a shared multi-tenant platform imposes no per-tenant limits on disk I/O or CPU consumption at all, letting any single tenant's batch job consume as much of the shared hardware's resources as it wants. What does this represent?
This is a missed opportunity to prevent a noisy neighbor problem, since imposing per-tenant resource quotas would prevent one tenant's heavy workload from degrading performance for every other tenant sharing the same hardware. A cache eviction policy is an unrelated concept about discarded cache entries. This no-per-tenant-limits pattern is exactly the kind of shared-resource risk a reviewer flags once multiple tenants are running on the same physical hardware.
4 / 5
An incident report shows dozens of unrelated tenants experienced severely degraded performance for hours, because one tenant's batch job consumed nearly all the shared hardware's disk I/O, and the platform had no per-tenant quota in place to cap that consumption. What practice would prevent this?
Imposing per-tenant resource quotas and I/O throttling caps one tenant's heavy workload before it can degrade performance for every other tenant sharing the same hardware. Continuing to impose no per-tenant limits on disk I/O or CPU consumption regardless of how many unrelated tenants experience degraded performance as a result is exactly what caused the widespread degradation described in this incident. This impose-per-tenant-quotas approach is the standard fix once a noisy neighbor problem is confirmed to be degrading performance for other tenants.
5 / 5
During a PR review, a teammate asks why the team imposes per-tenant resource quotas instead of simply giving every tenant dedicated, non-shared hardware to avoid the noisy neighbor problem entirely. What is the reasoning?
Per-tenant resource quotas keep the cost efficiency of shared infrastructure while still preventing one tenant's workload from degrading performance for others, whereas giving every tenant fully dedicated hardware also eliminates the noisy neighbor problem but gives up the cost and utilization benefits that sharing infrastructure across many tenants was meant to provide in the first place. This is exactly why per-tenant quotas are the standard fix for shared multi-tenant platforms, while dedicated hardware remains reserved for tenants whose needs justify that cost.