Per-workload host services connections

[autodidaddict]

Each workload is now given its own unique/private host services connection. These connections are created using the following order of configuration (1 being the highest priority)

1. host services configuration on the deployment request
2. host services configuration supplied for the entire node
3. configuration copied from the node's control connection

[autodidaddict]

It looks like the tests are failing in and around the use of host services (no surprise there). However, when doing manual tests it all works, so I'm wondering if there's just something that needs to be updated in the proxy or fixtures?

[autodidaddict]

The debug line showing the per-workload host services connection:

[INF] 2024-06-08 17:23:11 - Established host services connection for workload workload_name=echofunctionjs url=nats://127.0.0.1:4222

./nex devrun ../examples/v8/echofunction/src/echofunction.js --trigger_subject=foo.trigger --type=v8
Reusing existing issuer account key: /home/kevin/.nex/issuer.nk
Reusing existing publisher xkey: /home/kevin/.nex/publisher.xk
🚀 Workload 'echofunctionjs' accepted. You can now refer to this workload with ID: cpicmsn52omib90596g0 on node NC27GEVVHI6VTT6MOAH64VGZ5VMSYUKB5CWAELUY2SEC3QR4W5ZPJ4JE
kevin in wopr in nex/nex on  feat/perworkloadhsconfig via 🐹 v1.22.2 
❯ nats req 'foo.trigger' 'howdy'
17:23:49 Sending request on "foo.trigger"
17:23:49 Received with rtt 3.293876ms
{"triggered_on":"foo.trigger","payload":"howdy"}