Emit more breadcrumbs in `@sentry/node`

### Description

We should check all our existing Node instrumentation to see if we can easily emit breadcrumbs as well.

It is important not to tie this to spans, as these have no data when not sampled. Instead, where possible we need to use hooks in instrumentation to create breadcrumbs.

```[tasklist]
### Tasks
- [ ] genericPool
- [ ] kafka
- [ ] mongo
- [ ] mongoose
- [ ] mysql
- [ ] mysql2
- [ ] postgres
- [ ] prisma
- [ ] redis
- [ ] fs
```

Start by reviewing the existing @sentry/node instrumentation for genericPool, kafka, mongo, mongoose, mysql, mysql2, postgres, prisma, redis, and fs. Identify the available instrumentation hooks rather than relying on spans, then verify that each applicable integration emits breadcrumbs when unsampled. Done means the task list is addressed with breadcrumbs produced independently of span data.

Written by the indexing model from the issue text.