Repository Issue Activity (beta)

#968 Implement OTEL trace format replay

#980 [Bug]: Non-streaming audio transcription handler parses standard {"text": ...} responses as chat completions

#979 [Bug] gpt-oss benchmarks silently error out: unconditional ignore_eos is incompatible with Harmony chat completions

#613 Worker marks completed/successful when final response has no usable data

#969 The "plot" output kind lacks documentation

#940 Support structured chat content for google/translategemma-12b-it

#704 Display benchmark progress in non-interactive consoles

#966 DAG support for agentic conversation flows

#967 Implement WEKA trace format

#965 Agentic trace replay capabilities

#722 Turn and Conversation Metrics Plan

#719 --output-extras arg documented, but not implemented?

#86 Enable vLLM scheduler specific Metrics

#832 Websocket backend metrics

#834 Configurable backend RTF for realtime transcription

#833 Audio transcription metrics (RTF & RTFx)

#939 'guidellm benchmark from-file' crashes on console output in pip release 0.7.1

#588 --max-requests with --rate-type=sweep causes first stage to run all requests synchronously, resulting in extremely long runtimes

#937 CLI: 'benchmark' command help text is misleading

#761 GuideLLM 0.6.0 HTML report displays empty metrics despite valid benchmark data

#933 Inconsistent and limited handling of `--output` lists

#925 `guidellm benchmark from-file` crashes with `ValidationError` when output includes `console` format

#583 Display graphs generation as images

#604 Stop multi-rate benchmarks after first failure threshold

#924 Feature Request: Compare metrics across different benchmarks support