Alerting: support sending alerts to Alertmanagers requiring sigv4 authentication, like AMP

<!-- Please use this template to create your bug report. By providing as much info as possible you help us understand the issue, reproduce it and resolve it for you quicker. Therefore take a couple of extra minutes to make sure you have provided all info needed. PROTIP: record your screen and attach it as a gif to showcase the issue. - Questions should be posted to: https://community.grafana.com - Use query inspector to troubleshoot issues: https://bit.ly/2XNF6YS - How to record and attach gif: https://bit.ly/2Mi8T6K --> **What happened**: Added Amazon Managed Prometheus Alertmanager as an external alertmanager data source in Grafana. Tried sending Grafana based alerts to AMP Alertmanager but received a 403 error: ``` logger=ngalert.sender.external-alertmanager t=2023-03-02T18:38:03.676168206Z level=error alertmanager=https://aps-workspaces.eu-west-1.amazonaws.com/workspaces/ws-WORKSPACE-ID/alertmanager/api/v2/alerts count=9 msg="Error sending alert" err="bad response status 403 Forbidden" ``` Have tested the endpoint using `awscurl` and confirmed that an alert can be sent manually. The Alertmanager datasource is working. Health checks pass, alerts defined in alertmanager are listed and existing alerts via the groups tab are listed. **What you expected to happen**: Alert to be sent to AMP Alertmanager. **How to reproduce it (as minimally and precisely as possible)**: Setup AMP Alertmanager as an external alertmanager data source using sigv4 auth. Configure Grafana alerts to be sent to AMP Alertmanager. **Anything else we need to know?**: **Environment**: - Grafana version: 9.4.2 - Data source type & version: Alertmanager - OS Grafana is installed on: Deployed via Grafana Helm chart