excise unicode-byte-truncate usage, in favour of char-count-based truncation

In https://github.com/elastic/apm-nodejs-http-client/pull/162 we moved the http client away from unicode-byte-truncate for (a) conformance with other agents and (b) perf. For perf we should look at the current uses of unicode-byte-truncate in the agent code and see if there are perf wins there. For example, the following in parsers.js might be really slow in some cases.

```js
if (body.length > _MAX_HTTP_BODY_CHARS) {
body = truncate(body, _MAX_HTTP_BODY_CHARS)
```