new DependencyImpact library

[lzehl]

Controlled terminology to describe the impact dependencies being off-line or otherwise unavailable.

Related schema issue: openMetadataInitiative/openMINDS_controlledTerms#513

Suggested terms (for now):

• complete outage: the dependent becomes entirely nonfunctional.
• critical function loss: essential or mission-critical features stop working.
• non-critical function loss: secondary or optional features stop working while core functions continue.
• reduced performance: the dependent remains functional but responds slower or with higher latency.
• data unavailability: the dependent cannot retrieve any required data, causing operations to halt or fail.
• data staleness: the dependent must rely on cached or outdated data because fresh data cannot be retrieved.
• fallback mode activation: the dependent switches to an intentionally designed degraded or alternate operating mode.
• queue build-up: requests or tasks accumulate because normal processing cannot proceed.
• error propagation: the dependent emits errors that surface to users or downstream systems.

[lzehl]

revised definitions:

• complete outage: A [system|thing] becomes entirely nonfunctional if the thing it depends on is unavailable.
• critical function loss: Essential or mission-critical features of a [system|thing] stop working if the thing it depends on is unavailable.
• non-critical function loss: Secondary or optional features of a [system|thing] stop working while core functions continue if the thing it depends on is unavailable.
• reduced performance: A [system|thing] remains functional but responds slower or with higher latency if the thing it depends on is unavailable.
• data unavailability: A [system|thing] cannot retrieve any required data, causing operations to halt or fail, if the thing it depends on is unavailable.
• data staleness: A [system|thing] must rely on cached or outdated data because fresh data cannot be retrieved if the thing it depends on is unavailable.
• fallback mode activation: A [system|thing] switches to an intentionally designed degraded or alternate operating mode if the thing it depends on is unavailable.
• queue build-up: Requests or tasks directed to a [system|thing] accumulate because normal processing cannot proceed if the thing it depends on is unavailable.
• error propagation: A [system|thing] emits errors that surface to users or downstream systems if the thing it depends on is unavailable.

[apdavison]

"responds slower" -> "responds more slowly"

[lzehl]

maybe add specifally data loss and data corruption to the list