[core.logging] Fix calculating payload bytes for arrays and zlib streams.

[lukeelmers]

Closes #90457

Summary

This PR fixes 2 bugs related to how we calculate payload bytes for the purposes of http response logging which was introduced in #87939:

1. Ensures object responses that are arrays get calculated (in [core.logging] Fix issue where logs fail to calculate size of gunzip streams. #90353 we fixed a bug that was breaking Fleet, but forgot to consider array payload calculations)
2. Ensures that we calculate payload sizes for the zlib streams which are used in Fleet and SIEM.

Example logs with calculated zlib payloads

Reviewing / Testing

It is recommended to preview the diff with "hide whitespace changes" enabled.

You need to enable http.server.response logs to test this. Here's the config I've been testing with:

logging:
  events: # add this section to test legacy logs and compare
    request: "*"
    response: "*"

  appenders:
    console:
      type: console
      layout:
        type: pattern
        highlight: true
        pattern: "[%date][%level][%logger] %message"

  root:
    appenders: [console, default]
    level: warn

  loggers:
    - name: http.server.response
      appenders: [console]
      level: debug

[elasticmachine]

Pinging @elastic/kibana-core (Team:Core)

[lukeelmers]

@elasticmachine merge upstream

[lukeelmers]

@elasticmachine merge upstream

[mshustov]

Why we need such a sophisticated check? Maybe duck-typing would be enough?

[lukeelmers]

Re-evaluating each of the typeguards, there's actually quite a bit of cleanup we can do -- for example, we're already checking isBoom earlier in the function so we don't need those checks duplicated in the typeguards, and testing reference equality of res.source and src isn't strictly necessary.

I went ahead and simplified each of them to take the source object and the variety string.

The one for zlib streams is still a little more complex as TS will complain if we just try to check variety === 'stream' && obj.bytesWritten, as obj is unknown, hence the typeof and null checks.

[lukeelmers]

@elasticmachine merge upstream

[lukeelmers]

@elasticmachine merge upstream

@restrry This is ready for another look whenever you get the chance.

[lukeelmers]

@elasticmachine merge upstream

[lukeelmers]

@elasticmachine merge upstream

[kibanamachine]

💚 Build Succeeded

• continuous-integration/kibana-ci/pull-request
• Commit: e26d877
• Storybooks Preview
• Documentation Changes

Metrics [docs]

Async chunks

Total size of all lazy-loaded chunks that will be downloaded as the user navigates the app

id	before	after	diff
triggersActionsUi	1.6MB	1.5MB	-23.9KB

Page load bundle

Size of the bundles that are downloaded on every page load. Target size is below 100kb

id	before	after	diff
triggersActionsUi	104.0KB	104.1KB	+82.0B

Unknown metric groups

async chunk count

id	before	after	diff
triggersActionsUi	41	42	+1

History

• 💔 Build #110176 failed daae50f
• 💚 Build #109232 succeeded 99988e9
• 💚 Build #108922 succeeded 1261ef1
• 💚 Build #108921 succeeded d32f396
• 💚 Build #108608 succeeded a4b3bca

To update your PR or re-run it, just comment with:
@elasticmachine merge upstream

cc @lukeelmers

[kibanamachine]

💚 Backport successful

✅ 7.12 / #93284
✅ 7.x / #93285

Successful backport PRs will be merged automatically after passing CI.