Update the `vercel` README with accurate information about the managed Vercel endpoint

Currently, the [README](https://github.com/elastic/integrations/tree/main/packages/vercel#readme) of the `vercel` ("Vercel OpenTelemetry Integration”) integration is inaccurate with regard to where the user can find the Vercel managed endpoint.

The endpoint seems to be available in the UI only for Observability Serverless projects at the moment (in the staging environment, but not yet in production). It is not surfaced on the UI for ECH deployments. The integration docs refer to the endpoint being available for both Serverless and ECH (cf. [Requirements](https://github.com/elastic/integrations/tree/main/packages/vercel#requirements)).

[Step 2: Gather the managed endpoint URL and API key from Elastic](https://www.elastic.co/docs/reference/integrations/vercel#step-2-gather-the-managed-endpoint-url-and-api-key-from-elastic) section refers users to look for the endpoint on the project/deployment’s **Manage** page. This is inaccurate, as currently the Vercel endpoint is only surfaced on the **Add data** page of a Serverless Observability project, after clicking **More** in the **Connect directly to the endpoint** section.

If users can use the endpoint on ECH deployment by forming it from the OpenTelemetry/managed OTLP endpoint (the `ingest` endpoint), then this should also be explicitly explained, including how to form the Vercel endpoint from the base `ingest` endpoint.

Open the vercel integration README and review the Requirements and “Step 2: Gather the managed endpoint URL and API key from Elastic” sections. Verify the documented Add data > More > Connect directly to the endpoint path for Serverless Observability projects, then clarify whether ECH users can form the Vercel endpoint from the managed OTLP ingest endpoint. Done means the README accurately reflects the current UI and supported deployment options.

Written by the indexing model from the issue text.