Add metrics to phishing warning page

**Background**

We are iterating the current version of the phishing warning page. Ideally, we improve how we communicate the risk to users, so they decide NOT to move forward visiting the deceptive site.

Before iterating, we should implement some metrics to have a baseline for the current usage and be able to measure if the new version is indeed effective or not.

**Acceptance Criteria**

1. Phishing warning page event is triggered when users who have opted in MetaMetrics are prompted to it.
2. Phishing warning page link(s) event is triggered when users who have opted in MetaMetrics click on:
2.1 Main CTA [Back to safety]
2.2 Continue to the suspicious site [Continue to the site]
2.3 External link to report an error in the flagging
2.4 External link to Zendesk [Learn More]
2.5 External link for lists providers

**Metrics Checklist**

- [ ] This PR observes the privacy policy defined in [MetaMetrics: A Refresher on MetaMask’s Analytics System](https://medium.com/metamask/metametrics-a-refresher-on-metamasks-analytics-system-78d725cb9a75)
- [ ] This PR does NOT add any new 3rd party data providers
- [ ] This PR does NOT collect token balances, IP addresses, wallet addresses, or private keys

Start by locating the phishing warning page entry point and any existing MetaMetrics instrumentation. Add events for displaying the page and clicking each listed link, then verify the acceptance criteria and privacy checklist without introducing new data providers or collecting the prohibited data.

Written by the indexing model from the issue text.