Metric alerts not firing despite clear threshold breaches across multiple metrics

### Environment

SaaS (https://sentry.io/)

### Steps to Reproduce

1. Create a Metric Alert rule for any performance metric (e.g., Transaction Duration p75, LCP, INP, etc.) on a project with active traffic.
2. Configure thresholds that are clearly exceeded based on the data visible in the alert's own graph and preview.
3. Wait for multiple evaluation cycles to pass.
4. Observe that the alert status remains Resolved and no notifications are sent, despite the graph on the alert detail page showing values well within the critical zone.

Looks similar to #81724 where metric alerts stopped firing on SaaS without any rule modifications

### Expected Result

Alerts should transition to Critical or Warning status and send notifications when the monitored metric exceeds the configured threshold, as shown on the alert's own graph.

### Actual Result

All metric alert rules remain in Resolved status indefinitely. "No alerts triggered during this time" is displayed even though the graphs on the same pages show sustained threshold breaches over multiple times.
I set stricter conditions and tried different filters, but the status remained unchanged.

### Product Area

Alerts

### Link

_No response_

### DSN

_No response_

### Version

Sentry 25.10.0

No source file or test is identified in the issue. Start by reproducing a metric alert in the Sentry SaaS environment and compare the alert graph with its evaluation status across multiple cycles. Done means sustained threshold breaches transition the alert to Critical or Warning and send notifications.

Written by the indexing model from the issue text.