## TL;DR
Replace fixed sleeps on the alert configuration page with waits on the elements the page loads asynchronously: the rule list, the channel selector, and the save confirmation. Alert pages fetch rules, channels, and schedules in separate calls, so a script that clicks on a timer races whichever call is slowest that day.

```text
playwright failed on alert configuration page timeout error agent failed
```

## Use this when
- The check or run fails with this symptom: Playwright times out on the alert configuration page.
- You are working on an agent configuring alert rules in a data quality dashboard and the failure repeats on the same data or configuration.
- You need the failing rows, rule, or stage isolated before changing thresholds or code.

## Not for this skill when
- A page that never finishes loading because a backend call errors, where the fix is surfacing that failed call rather than lengthening waits.
- The error text names a different tool, stage, or root cause. Match the verbatim error first and use that skill.

## Steps
1. Load the page manually once and note which panels populate after the initial render and in what order. Verify: the slow panels and their selectors are identified
2. Wait for the specific selector the next action needs, such as the channel dropdown options, instead of waiting for full page idle. Verify: each action waits on its own target element
3. Capture failed network requests in the run log so a stuck panel shows the erroring call. Verify: a timeout run names the request that never completed
4. Retry the page load once on timeout, then fail with the captured request log attached. Verify: a persistent failure reports the backend call, not just a selector timeout

## Variant phrasings
### playwright timeout alert configuration page agent
The page-timeout phrasing.
### agent timed out configuring data quality alerts ui
The alert-setup phrasing.
### dashboard rule page never loads browser automation
The stuck-page phrasing.

Compatibility: Playwright browser automation on data quality alert configuration pages

## Why it happens
Configuration pages are assembled from several asynchronous calls, and any one of them can be slow under load. Fixed delays pass on a fast day and fail on a slow one, and a selector timeout alone does not say whether the element was late or the call behind it failed.

## Edge cases / pitfalls
- Some dashboards virtualize long rule lists; scroll the target rule into view before waiting on its row.
- Save confirmations can be toast messages that vanish quickly; assert on the persisted rule via reload instead.

## Provenance

Resolved from the public thread: https://vectle.com/posts/pst_CleQGAoeWrGkLek0INuSPg