feat: Improve agent startup

[ptang-nr]

Provide better support for scenarios where agent startup can be interrupted before RUM call completes successfully and feature session modes are saved into local storage (where applicable).

Overview

In this PR,

• the agent startup avoids defaulting and locking a new session into OFF mode for Session Trace, Session Replay, and/or Browser Logging when the first RUM call in a session is incomplete (for example, due to redirects or early navigations).
• additionally when the RUM call completes, the agent will now cache the RUM response in the session/local storage such that the same feature modes can be reused for the duration of the session.
  • RUM call sends with fsh = 1, when it is the first RUM call in the session and there are no cached RUM response
  • RUM call sends with fsh = 0, when a cached RUM response is detected
• if a cached RUM response is present for a given session, the agent will use the cached response to continue startup

Related Issue(s)

https://new-relic.atlassian.net/browse/NR-509779

Testing

Extended integration tests in affected parts of the agent.
Manually tested scenario to ensure RUM response is cached in the session.

[github-actions]

Supportability Metric Usage Changes:

No matching changes found

Supportability Metrics .md File Changes:

supportability_metrics.md was changed? false

New supportability metrics require changes to supportability_metrics.md, as well as a new PR to Angler. Please ensure an Angler PR is created and linked to this PR.

[github-actions]

Asset Size Report

Merging this pull request will result in the following asset size changes:

Agent	Asset	Previous Size	New Size	Diff
lite	loader	26.79 kB / 10.24 kB (gzip)	26.86 kB / 10.26 kB (gzip)	0.23% / 0.16% (gzip)
lite	async-chunk	58.92 kB / 19.76 kB (gzip)	59.62 kB / 20.05 kB (gzip)	1.18% / 1.49% (gzip)
pro	loader	66.61 kB / 23.46 kB (gzip)	66.89 kB / 23.53 kB (gzip)	0.41% / 0.27% (gzip)
pro	async-chunk	96.39 kB / 30.65 kB (gzip)	97.17 kB / 30.83 kB (gzip)	0.81% / 0.57% (gzip)
spa	loader	69.01 kB / 24.24 kB (gzip)	69.28 kB / 24.3 kB (gzip)	0.4% / 0.25% (gzip)
spa	async-chunk	108.54 kB / 33.76 kB (gzip)	109.32 kB / 34.02 kB (gzip)	0.72% / 0.76% (gzip)

[codecov]

Codecov Report

❌ Patch coverage is 89.74359% with 16 lines in your changes missing coverage. Please review.
✅ Project coverage is 90.07%. Comparing base (f7df627) to head (6076551).
⚠️ Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
src/common/harvest/send.js	88.37%	8 Missing and 2 partials ⚠️
src/features/session_replay/aggregate/index.js	85.00%	3 Missing ⚠️
src/features/session_trace/aggregate/index.js	85.71%	2 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1760      +/-   ##
==========================================
+ Coverage   89.54%   90.07%   +0.53%     
==========================================
  Files         213      214       +1     
  Lines        7864     7852      -12     
  Branches     1657     1669      +12     
==========================================
+ Hits         7042     7073      +31     
+ Misses        743      709      -34     
+ Partials       79       70       -9     

Flag	Coverage Δ
unit-tests	86.03% <89.74%> (+1.06%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[metal-messiah]

Just some general WIP comments, looking good!

[github-actions]

Last ran on June 16, 2026 13:01:32 CDT
Checking merge of (a58738a) into main (3fd403e)

[metal-messiah]

This looks good and makes sense. I assume app metadata is unneeded for the cache, but just want to double check that above ⬆️ before merging.