Deploy on Node 16

[chris48s]

📋 Description

Node 16 has been generally available for ~6 months now and includes a number of performance improvements we may benefit from. We've had the full test suite consistently passing for some time and we're now starting to see dependency upgrades which require us to make the jump. I see no reason not to 🚀 in production in the near future.

[chris48s]

e.g: #7030 (comment)

[PyvesB]

Worth noting that I did notice a few deprecation warnings when running some of our NPM scripts with Node 16. They are by no means a blocker, but it would be good to clean them up at some point.

[chris48s]

So I had a look at the deprecations.
As far as I can tell, although DeprecationWarnings get thrown at a few different points when running the test suite, they're all instances of the same error:

(node:21960) [DEP0111] DeprecationWarning: Access to process.binding('http_parser') is deprecated.
(Use `node --trace-deprecation ...` to show where the warning was created)

Unfortunately, this isn't an error which is being thrown by code in our codebase. It is http-deceiver which is a transitive dependency. We depend on @shields_io/camp which depends on spdy which depends on http-deceiver

├─┬ @shields_io/[email protected]
│ ├─┬ [email protected]
│ │ ├── [email protected]

I did have a look to see if we can fix this by updating the requirements on our scoutcamp fork, but this is an open issue on http-deceiver: spdy-http2/http-deceiver#6
In general, this entire branch of the dependency tree is somewhat unloved.

• https://www.npmjs.com/package/spdy last had a release 2 years ago
• https://www.npmjs.com/package/http-deceiver last had a release 5 years ago

so I think we're probably not going to resolve this any time soon and I don't think we should block upgrading to node 16 on it. More widely, moving away from ScoutCamp is already an objective for us and the fact that staying with it may block a future node upgrade bumps it up the priority order.