feat: collect-and-flush-analytics-evaluation-events

[Zaimwa9]

Thanks for submitting a PR! Please check the boxes below:

• I have read the Contributing Guide.
• I have added information to docs/ if required so people know about the feature.
• I have filled in the "Changes" section below.
• I have filled in the "How did you test this code" section below.

Changes

Contributes to 6636

• Add evaluation analytics pipeline integration: buffer individual flag evaluation events and flush them to a configurable pipeline endpoint (evaluationAnalyticsConfig)
• Record event data per evaluation, as of now: flag key, value, enabled state, identity, traits at evaluation time, and timestamp
• Re-queue events on flush failure with buffer cap

Usage

// JS
flagsmith.init({
      environmentID: 'your_environment_key',
      evaluationAnalyticsConfig: {
          analyticsServerUrl: 'https://analytics.flagsmith.com/',
          maxBuffer: 500,
          flushInterval: 5000,
      },
  });
  
  // React
    <FlagsmithProvider
    flagsmith={flagsmith}
    options={{
      environmentID: 'your_environment_key',
      evaluationAnalyticsConfig: {
        analyticsServerUrl: 'https://analytics.flagsmith.com/',
      },
    }}
  >
    {children}
  </FlagsmithProvider>

Bonus

• Added the pull request template

Improvements for V2

• Retry with exponential backoff: currently re-queues on failure but retries at the same interval
• Flush on session end use visibilitychange / beforeunload + navigator.sendBeacon() to flush buffered events before tab close (the size of the payload might be too heavy versus the simple count of previous analytics to be stored in localStorage)
• Persist unflushed events: persist buffer to so events survive page refreshes
• Extract PipelineAnalytics class: isolate analytics logic (buffer, flush, intervals) from the core to gain in flexibility without sacrificing maintainability
• destroy() method (cf openFeature): proper teardown to prevent memory leaks in SSR/Next.js (pre-existing gap in all interval-based feature in SDK atm)

How did you test this code?

• Added dedicated tests

[talissoncosta]

I see that 'v1/analytics/batch' and 'analytics/flags' appear multiple times. Would be better to add it as const ?

Not blocking, just a suggestion

[talissoncosta]

Great work @Zaimwa9

[Zaimwa9]

This is temporary to allow testing without redeploying the rust service. Will be reverted to null once made optional there