Playwright: expose wait-for-grafana timeout (or raise default to 90s)

## Problem

In `playwright.yml` (v7.0.0), the `Wait for Grafana to start` step calls `grafana/plugin-actions/wait-for-grafana@v1.0.2` without passing a `timeout`, so it uses the action's hardcoded default of **60 seconds**.

On larger images (e.g., `grafana-enterprise:13.0.1`, `dev-preview-react19`) Grafana sometimes takes longer than 60s to become healthy on the shared runners — especially when the image hasn't been pulled recently. This causes intermittent fast-fail CI flakes before any test runs. Observed repeatedly on [grafana/business-table PR #51](https://github.com/grafana/business-table/pull/51) across multiple Grafana versions.

Example failure log:
\`\`\`
GRAFANA_TIMEOUT: 60
Timeout: 60 seconds
Waiting for server to respond with status code 200. Current status: 000
... (60s of 000) ...
##[error]Process completed with exit code 1.
\`\`\`

## Request

Either of these would fix it for plugin repos:

1. **Raise the default** passed to `wait-for-grafana` in `playwright.yml` to 90s. The action's internal default is 60 but the reusable workflow can set a safer floor.

2. **Expose a workflow input** on `playwright.yml` (and `ci.yml`) — something like `playwright-wait-for-grafana-timeout: 90` — so plugin repos can tune it themselves without waiting for a workflow release.

Option 2 is more flexible long-term. Option 1 is zero-config for everyone.

Happy to send a PR if either is welcome.