Allow expected values to be overridden

[nsleigh]

What would you like?
Allow the expected values to be overridden so that it is possible to moitor an existing disk with an issue and highlight when it gets worse.

Why?
In my case I have a disk in my NAS with "Reallocated Sector Count: 147 (expected 0); Reallocated Event Count: 47 (expected 0)" - to my knowledge it has been like this for years (I think it happened after a poor shutdown - but until smart-sniffer I never monitored). As per my understanding these issues are not inherently bad but are if they get worse.

Alternatives considered
As a workaround I can setup an automation to 'know' the error values and monitor for changes but I would like something simpler. Currently I use AlertTicker-Card for monitoring all the drive temperatures and it would be easy to add monitoring of the 'Attention needed' if it could be reset.

Additional context
N/A

[DAB-LABS]

Welcome to the operation, Agent @nsleigh! 🕵️‍♂️

Good news is that configurable alert thresholds have been on our roadmap, and your use case is exactly why. A drive with 147 reallocated sectors that's been stable for years shouldn't be permanently screaming at you.

We're working on this now. The plan is to add a per-drive settings page in the integration's Configure dialog where you can see each monitored attribute, its current value, and set a threshold. The attention sensor would only trigger when the value exceeds your threshold instead of the default (zero for most attributes). So you'd set Reallocated Sector Count to 147, Reallocated Event Count to 47, and the drive would drop from YES to NO -- and only alert you again if the counts actually increase beyond those numbers.

🫡

~DAB

---

📋 Field Brief: Configurable Alert Thresholds

Subject: No way to set accepted baseline values for known-degraded drives; Attention Needed sensor permanently triggers on stable historical damage
Classification: Feature Request (roadmap item)

Situation: User has a NAS drive with 147 reallocated sectors and 47 reallocated events that have been stable for years. The attention evaluation logic in attention.py compares critical ATA attributes against a hardcoded threshold of 0. Any non-zero value triggers YES/critical. There is no mechanism to override the expected value on a per-drive or per-attribute basis. The user cannot distinguish "known stable damage" from "new active degradation."

Impact:

User type	Impact
Users with known-degraded but stable drives	Attention sensor permanently stuck on YES; sensor becomes meaningless
Users monitoring attention via automations	Cannot use Attention Needed as a trigger because it never clears
Users with AlertTicker-Card or similar dashboards	Cannot integrate attention monitoring alongside temperature

Root Cause: evaluate_attention() in attention.py uses hardcoded thresholds:

Attribute type	Current threshold	Comparison
ATA critical (reallocated, pending, uncorrectable)	0	> 0 triggers alert
ATA warning (reallocated events, spin retry)	0	> 0 triggers alert
Command Timeout	100	> 100 triggers alert
SSD/NVMe wear	90%	>= 90% triggers alert

No mechanism exists to override these per drive or per attribute.

Resolution: Add configurable per-drive thresholds via the HA integration's options flow. Plan doc: docs/internal/plans/plan-configurable-alert-thresholds.md.

Component	Change
attention.py	Add thresholds parameter to evaluate_attention(), compare against user-set values instead of hardcoded defaults
config_flow.py	Add two-step options flow: drive picker, then threshold editor showing attribute name, current value, and editable threshold
sensor.py	Pass thresholds from config_entry.options to evaluate_attention()
coordinator.py	Same -- pass thresholds for notification handling
const.py, strings.json	New constants and UI strings

Key design decisions:

Decision	Choice	Rationale
Where to store thresholds	config_entry.options	HA-standard, persists across restarts, backs up with HA
When to configure	Post-setup, via Configure dialog	Integration starts with defaults; user adjusts after seeing their drives
Scope	Per-drive, per-attribute	Maximum flexibility; each drive has its own situation
Default behavior	Identical to current (threshold=0)	Zero change for existing users who don't touch settings

Estimated effort: ~3.5 hours (mostly options flow UX)

Roadmap:

Feature	Status
Configurable alert thresholds via options flow	In development
Auto-accept current values (one-click baseline)	Deferred (UX enhancement)
Global default thresholds	Deferred

---