feat: OpenAI Integration

[PROFeNoM]

Description

TBC

GL reviewing

Reviewer checklist

• [ ] Test coverage seems ok.
• [ ] Appropriate labels assigned.

[pr-commenter[bot]]

Benchmarks

Benchmark execution time: 2024-07-19 22:47:29

Comparing candidate commit 841220b85e687f167512be07f2103b613994e5b7 in PR branch alex/feat/openai with baseline commit 5fe1c5ae9ff97ecc011135fdd0ddddb0452c1cbc in branch master.

Found 3 performance improvements and 48 performance regressions! Performance is the same for 127 metrics, 0 unstable metrics.

scenario:ComposerTelemetryBench/benchTelemetryParsing

• 🟥 mem_peak [+107.144KB; +107.144KB] or [+3.419%; +3.419%]

scenario:ContextPropagationBench/benchExtractHeaders128Bit

• 🟥 mem_peak [+107.144KB; +107.144KB] or [+3.418%; +3.418%]

scenario:ContextPropagationBench/benchExtractHeaders64Bit

• 🟥 mem_peak [+107.144KB; +107.144KB] or [+3.418%; +3.418%]

scenario:ContextPropagationBench/benchExtractTraceContext128Bit

• 🟥 mem_peak [+107.144KB; +107.144KB] or [+3.418%; +3.418%]

scenario:ContextPropagationBench/benchExtractTraceContext64Bit

• 🟥 mem_peak [+107.144KB; +107.144KB] or [+3.418%; +3.418%]

scenario:ContextPropagationBench/benchInject128Bit

• 🟥 mem_peak [+107.144KB; +107.144KB] or [+3.418%; +3.418%]

scenario:ContextPropagationBench/benchInject64Bit

• 🟥 mem_peak [+107.144KB; +107.144KB] or [+3.418%; +3.418%]

scenario:ContextPropagationBench/benchInject64Bit-opcache

• 🟥 execution_time [+186.587ns; +415.413ns] or [+2.535%; +5.645%]

scenario:EmptyFileBench/benchEmptyFileBaseline

• 🟥 mem_peak [+246.024KB; +246.024KB] or [+5.721%; +5.721%]

scenario:EmptyFileBench/benchEmptyFileOverhead

• 🟥 mem_peak [+246.024KB; +246.024KB] or [+5.721%; +5.721%]

scenario:HookBench/benchHookOverheadInstallHookOnFunction

• 🟥 mem_peak [+107.144KB; +107.144KB] or [+3.418%; +3.418%]

scenario:HookBench/benchHookOverheadInstallHookOnMethod

• 🟥 mem_peak [+107.144KB; +107.144KB] or [+3.418%; +3.418%]

scenario:HookBench/benchHookOverheadTraceFunction

• 🟥 mem_peak [+107.480KB; +107.480KB] or [+3.282%; +3.282%]

scenario:HookBench/benchHookOverheadTraceMethod

• 🟥 mem_peak [+107.479KB; +107.482KB] or [+3.219%; +3.219%]

scenario:HookBench/benchWithoutHook

• 🟥 mem_peak [+107.144KB; +107.144KB] or [+3.419%; +3.419%]

scenario:LaravelBench/benchLaravelBaseline

• 🟥 mem_peak [+246.056KB; +246.056KB] or [+5.720%; +5.720%]

scenario:LaravelBench/benchLaravelOverhead

• 🟥 mem_peak [+246.056KB; +246.056KB] or [+5.720%; +5.720%]

scenario:LogsInjectionBench/benchLogsInfoBaseline

• 🟥 mem_peak [+135.288KB; +135.288KB] or [+3.859%; +3.859%]

scenario:LogsInjectionBench/benchLogsInfoInjection

• 🟥 mem_peak [+139.264KB; +139.264KB] or [+3.949%; +3.949%]

scenario:LogsInjectionBench/benchLogsNullBaseline

• 🟥 mem_peak [+135.288KB; +135.288KB] or [+3.859%; +3.859%]

scenario:LogsInjectionBench/benchLogsNullInjection

• 🟥 mem_peak [+139.320KB; +139.320KB] or [+3.947%; +3.947%]

scenario:MessagePackSerializationBench/benchMessagePackSerialization

• 🟥 mem_peak [+107.144KB; +107.144KB] or [+3.321%; +3.321%]

scenario:PDOBench/benchPDOBaseline

• 🟥 mem_peak [+71.632KB; +71.632KB] or [+2.047%; +2.047%]

scenario:PDOBench/benchPDOOverhead

• 🟥 execution_time [+7.473µs; +16.463µs] or [+2.739%; +6.034%]
• 🟥 mem_peak [+135.318KB; +135.321KB] or [+3.217%; +3.217%]

scenario:PDOBench/benchPDOOverhead-opcache

• 🟥 mem_peak [+125.248KB; +125.248KB] or [+5.964%; +5.964%]
• 🟩 execution_time [-10.465µs; -8.390µs] or [-3.512%; -2.815%]

scenario:PDOBench/benchPDOOverheadWithDBM

• 🟥 mem_peak [+135.318KB; +135.321KB] or [+3.172%; +3.172%]

scenario:PDOBench/benchPDOOverheadWithDBM-opcache

• 🟥 mem_peak [+125.247KB; +125.250KB] or [+5.800%; +5.800%]

scenario:PHPRedisBench/benchRedisOverhead

• 🟥 mem_peak [+135.335KB; +135.337KB] or [+2.675%; +2.675%]

scenario:PHPRedisBench/benchRedisOverhead-opcache

• 🟥 mem_peak [+125.263KB; +125.266KB] or [+4.185%; +4.185%]

scenario:SamplingRuleMatchingBench/benchGlobMatching1

• 🟥 mem_peak [+107.144KB; +107.144KB] or [+3.419%; +3.419%]

scenario:SamplingRuleMatchingBench/benchGlobMatching1-opcache

• 🟩 execution_time [-149.145ns; -71.655ns] or [-5.730%; -2.753%]

scenario:SamplingRuleMatchingBench/benchGlobMatching2

• 🟥 mem_peak [+107.144KB; +107.144KB] or [+3.419%; +3.419%]

scenario:SamplingRuleMatchingBench/benchGlobMatching3

• 🟥 mem_peak [+107.144KB; +107.144KB] or [+3.419%; +3.419%]

scenario:SamplingRuleMatchingBench/benchGlobMatching4

• 🟥 mem_peak [+107.144KB; +107.144KB] or [+3.419%; +3.419%]

scenario:SamplingRuleMatchingBench/benchRegexMatching1

• 🟥 mem_peak [+107.144KB; +107.144KB] or [+3.419%; +3.419%]

scenario:SamplingRuleMatchingBench/benchRegexMatching2

• 🟥 mem_peak [+107.144KB; +107.144KB] or [+3.419%; +3.419%]

scenario:SamplingRuleMatchingBench/benchRegexMatching3

• 🟥 mem_peak [+107.144KB; +107.144KB] or [+3.419%; +3.419%]

scenario:SamplingRuleMatchingBench/benchRegexMatching3-opcache

• 🟥 execution_time [+202.523ns; +405.877ns] or [+3.177%; +6.366%]

scenario:SamplingRuleMatchingBench/benchRegexMatching4

• 🟥 execution_time [+175.310ns; +401.690ns] or [+2.704%; +6.196%]
• 🟥 mem_peak [+107.144KB; +107.144KB] or [+3.419%; +3.419%]

scenario:SamplingRuleMatchingBench/benchRegexMatching4-opcache

• 🟥 execution_time [+130.057ns; +358.743ns] or [+2.028%; +5.593%]

scenario:SymfonyBench/benchSymfonyBaseline

• 🟥 mem_peak [+246.056KB; +246.056KB] or [+5.720%; +5.720%]

scenario:SymfonyBench/benchSymfonyOverhead

• 🟥 mem_peak [+246.056KB; +246.056KB] or [+5.720%; +5.720%]

scenario:TraceAnnotationsBench/benchTraceAnnotationOverhead

• 🟥 mem_peak [+107.480KB; +107.484KB] or [+3.211%; +3.211%]

scenario:TraceFlushBench/benchFlushTrace

• 🟥 mem_peak [+135.256KB; +135.256KB] or [+3.699%; +3.699%]

scenario:TraceSerializationBench/benchSerializeTrace

• 🟥 mem_peak [+107.144KB; +107.144KB] or [+3.418%; +3.418%]
• 🟩 execution_time [-11.128µs; -8.372µs] or [-5.677%; -4.272%]

scenario:WordPressBench/benchWordPressBaseline

• 🟥 mem_peak [+250.216KB; +250.216KB] or [+3.717%; +3.717%]

scenario:WordPressBench/benchWordPressOverhead

• 🟥 mem_peak [+246.024KB; +246.024KB] or [+5.712%; +5.712%]

[bwoebi]

I'd like to discuss the DatadogLogger, and would like to know if this needs to be done this way.

At a glance, I'm not particularly fond of it. I don't mind having logging. But I think the extension should expose an API to log messages, rather than having a custom logger written in PHP. Which then get sent to the right location, can be forwarded to the sidecar, possibly for logging telemetry and such in future.

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 94.68713% with 45 lines in your changes missing coverage. Please review.

Project coverage is 78.62%. Comparing base (5fe1c5a) to head (841220b).

Additional details and impacted files

@@             Coverage Diff              @@
##             master    #2685      +/-   ##
============================================
+ Coverage     77.97%   78.62%   +0.65%     
- Complexity     2216     2459     +243     
============================================
  Files           228      230       +2     
  Lines         26951    27794     +843     
  Branches        989      989              
============================================
+ Hits          21015    21853     +838     
- Misses         5410     5415       +5     
  Partials        526      526              

Flag	Coverage Δ
appsec-extension	69.19% <ø> (ø)
tracer-extension	78.82% <100.00%> (+<0.01%)	:arrow_up:
tracer-php	82.13% <94.64%> (+1.57%)	:arrow_up:

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

Files	Coverage Δ
ext/configuration.h	100.00% <ø> (ø)
ext/git.c	74.29% <100.00%> (+0.12%)	:arrow_up:
ext/integrations/integrations.c	91.25% <100.00%> (+0.14%)	:arrow_up:
ext/startup_logging.c	86.06% <ø> (ø)
src/api/Log/InterpolateTrait.php	77.77% <100.00%> (+77.77%)	:arrow_up:
src/bridge/_files_tracer.php	100.00% <100.00%> (ø)
src/api/Log/DatadogLogger.php	80.00% <80.00%> (ø)
.../DDTrace/Integrations/OpenAI/OpenAIIntegration.php	96.50% <96.50%> (ø)

... and 1 file with indirect coverage changes

---

Continue to review full report in Codecov by Sentry.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update 5fe1c5a...841220b. Read the comment docs.