Skip to content

Perf regression: benchmark outside its baseline band #275

Open
Open
Perf regression: benchmark outside its baseline band#275

Description

@github-actions
github-actions
bot
opened on Sep 20, 2026

Perf regression vs Tests/Verbara.Sdk.Benchmarks/baseline.json

Mode: OBSERVING (reports, does not fail the job — §2.5)

Band breaches (15)

  • [Verbara.Sdk.Benchmarks.AmiProtocolReaderBenchmark.Parse100EventBatch] FASTER than baseline by 25.04%, outside the +/-10% band by 15.04 pp. baseline 88642 ns, observed 66442.74 ns.
  • [Verbara.Sdk.Benchmarks.AmiProtocolReaderBenchmark.ParseResponse] FASTER than baseline by 28.89%, outside the +/-25% band by 3.89 pp. baseline 981.16 ns, observed 697.68 ns.
  • [Verbara.Sdk.Benchmarks.AmiProtocolReaderBenchmark.ParseSingleEvent] FASTER than baseline by 28.76%, outside the +/-25% band by 3.76 pp. baseline 1484 ns, observed 1057.27 ns.
  • [Verbara.Sdk.Benchmarks.AriJsonBenchmark.Deserialize100Channels] FASTER than baseline by 26.16%, outside the +/-15% band by 11.16 pp. baseline 64354 ns, observed 47519.57 ns.
  • [Verbara.Sdk.Benchmarks.AriJsonBenchmark.DeserializeBridge] FASTER than baseline by 30.99%, outside the +/-30% band by 0.99 pp. baseline 562.22 ns, observed 387.97 ns.
  • [Verbara.Sdk.Benchmarks.AriJsonBenchmark.DeserializeChannel] FASTER than baseline by 25.29%, outside the +/-20% band by 5.29 pp. baseline 666.79 ns, observed 498.17 ns.
  • [Verbara.Sdk.Benchmarks.AriParseEventBenchmark.ParseChannelDtmf] FASTER than baseline by 29.79%, outside the +/-25% band by 4.79 pp. baseline 1939 ns, observed 1361.40 ns.
  • [Verbara.Sdk.Benchmarks.AriParseEventBenchmark.ParseStasisStart] FASTER than baseline by 31.27%, outside the +/-25% band by 6.27 pp. baseline 3947 ns, observed 2712.65 ns.
  • [Verbara.Sdk.Benchmarks.AriParseEventBenchmark.ParseUnknownEvent] FASTER than baseline by 31.26%, outside the +/-25% band by 6.26 pp. baseline 547.36 ns, observed 376.27 ns.
  • [Verbara.Sdk.Benchmarks.ChannelManagerBenchmark.LookupByName] FASTER than baseline by 20.62%, outside the +/-15% band by 5.62 pp. baseline 14.69 ns, observed 11.66 ns.
  • [Verbara.Sdk.Benchmarks.ChannelManagerBenchmark.LookupByUniqueId] FASTER than baseline by 20.28%, outside the +/-15% band by 5.28 pp. baseline 12.01 ns, observed 9.57 ns.
  • [Verbara.Sdk.Benchmarks.ChannelManagerBenchmark.Update1000ChannelStates] FASTER than baseline by 26.38%, outside the +/-15% band by 11.38 pp. baseline 65925 ns, observed 48535.86 ns.
  • [Verbara.Sdk.Benchmarks.ObserverDispatchBenchmark.Dispatch_100Observers] FASTER than baseline by 26.05%, outside the +/-15% band by 11.05 pp. baseline 47.61 ns, observed 35.21 ns.
  • [Verbara.Sdk.Benchmarks.ObserverDispatchBenchmark.Dispatch_10Observers] FASTER than baseline by 28.52%, outside the +/-20% band by 8.52 pp. baseline 5.67 ns, observed 4.05 ns.
  • [Verbara.Sdk.Benchmarks.ObserverDispatchBenchmark.Dispatch_1Observer] FASTER than baseline by 22.95%, outside the +/-15% band by 7.95 pp. baseline 0.94 ns, observed 0.72 ns.

Measured

benchmark baseline (ns) observed (ns) delta band
AmiProtocolReaderBenchmark.Parse100EventBatch 88642 66442.74 -25.04% ±10% BREACH (FASTER)
AmiProtocolReaderBenchmark.ParseResponse 981.16 697.68 -28.89% ±25% BREACH (FASTER)
AmiProtocolReaderBenchmark.ParseSingleEvent 1484 1057.27 -28.76% ±25% BREACH (FASTER)
AriJsonBenchmark.Deserialize100Channels 64354 47519.57 -26.16% ±15% BREACH (FASTER)
AriJsonBenchmark.DeserializeBridge 562.22 387.97 -30.99% ±30% BREACH (FASTER)
AriJsonBenchmark.DeserializeChannel 666.79 498.17 -25.29% ±20% BREACH (FASTER)
AriJsonBenchmark.SerializeBridge 264.5 185.58 -29.84% ±35% ok
AriJsonBenchmark.SerializeChannel 363.01 238.29 -34.36% ±45% ok
AriParseEventBenchmark.ParseChannelDtmf 1939 1361.40 -29.79% ±25% BREACH (FASTER)
AriParseEventBenchmark.ParseStasisStart 3947 2712.65 -31.27% ±25% BREACH (FASTER)
AriParseEventBenchmark.ParseUnknownEvent 547.36 376.27 -31.26% ±25% BREACH (FASTER)
ChannelManagerBenchmark.Create1000Channels 353056 252470.89 -28.49% ±30% ok
ChannelManagerBenchmark.EnumerateByState 125710 91601.56 -27.13% ±30% ok
ChannelManagerBenchmark.LookupByName 14.69 11.66 -20.62% ±15% BREACH (FASTER)
ChannelManagerBenchmark.LookupByUniqueId 12.01 9.57 -20.28% ±15% BREACH (FASTER)
ChannelManagerBenchmark.Update1000ChannelStates 65925 48535.86 -26.38% ±15% BREACH (FASTER)
ObserverDispatchBenchmark.Dispatch_100Observers 47.61 35.21 -26.05% ±15% BREACH (FASTER)
ObserverDispatchBenchmark.Dispatch_10Observers 5.67 4.05 -28.52% ±20% BREACH (FASTER)
ObserverDispatchBenchmark.Dispatch_1Observer 0.94 0.72 -22.95% ±15% BREACH (FASTER)

Measured but not baselined (4)

Not a failure. Add a benchmarks entry once its hosted-runner spread is observed — see baseline.README.md.

  • Verbara.Sdk.Benchmarks.TurnDetectionBenchmark.EndOfUtteranceDecision(UtteranceSeconds: 1)
  • Verbara.Sdk.Benchmarks.TurnDetectionBenchmark.EndOfUtteranceDecision(UtteranceSeconds: 2)
  • Verbara.Sdk.Benchmarks.TurnDetectionBenchmark.EndOfUtteranceDecision(UtteranceSeconds: 4)
  • Verbara.Sdk.Benchmarks.TurnDetectionBenchmark.EndOfUtteranceDecision(UtteranceSeconds: 8)

Run: https://github.com/verbara/Verbara.Sdk/actions/runs/35501530838

Baselines move only by human-authored commit (ADR-0042 D6) — CI never writes back to
Tests/Verbara.Sdk.Benchmarks/baseline.json. If this is a real regression, fix the code.
If the hosted-runner fleet moved under us, re-baseline per
Tests/Verbara.Sdk.Benchmarks/baseline.README.md: a PR stating which benchmark moved, in
which direction, by how much, and why.

Activity

  1. github-actions commented on Sep 20, 2026

    @github-actions
    github-actionsbot
    on Sep 20, 2026 – with GitHub Actions
    Author

    Perf regression vs Tests/Verbara.Sdk.Benchmarks/baseline.json

    Mode: OBSERVING (reports, does not fail the job — §2.5)

    Band breaches (15)

    • [Verbara.Sdk.Benchmarks.AmiProtocolReaderBenchmark.Parse100EventBatch] FASTER than baseline by 37.05%, outside the +/-10% band by 27.05 pp. baseline 88642 ns, observed 55801.18 ns.
    • [Verbara.Sdk.Benchmarks.AmiProtocolReaderBenchmark.ParseResponse] FASTER than baseline by 36.83%, outside the +/-25% band by 11.83 pp. baseline 981.16 ns, observed 619.83 ns.
    • [Verbara.Sdk.Benchmarks.AmiProtocolReaderBenchmark.ParseSingleEvent] FASTER than baseline by 38.82%, outside the +/-25% band by 13.82 pp. baseline 1484 ns, observed 907.93 ns.
    • [Verbara.Sdk.Benchmarks.AriJsonBenchmark.Deserialize100Channels] FASTER than baseline by 41.54%, outside the +/-15% band by 26.54 pp. baseline 64354 ns, observed 37622.70 ns.
    • [Verbara.Sdk.Benchmarks.AriJsonBenchmark.DeserializeChannel] FASTER than baseline by 41.81%, outside the +/-20% band by 21.81 pp. baseline 666.79 ns, observed 388.01 ns.
    • [Verbara.Sdk.Benchmarks.AriParseEventBenchmark.ParseChannelDtmf] FASTER than baseline by 34.37%, outside the +/-25% band by 9.37 pp. baseline 1939 ns, observed 1272.63 ns.
    • [Verbara.Sdk.Benchmarks.AriParseEventBenchmark.ParseStasisStart] FASTER than baseline by 36.73%, outside the +/-25% band by 11.73 pp. baseline 3947 ns, observed 2497.34 ns.
    • [Verbara.Sdk.Benchmarks.AriParseEventBenchmark.ParseUnknownEvent] FASTER than baseline by 34.58%, outside the +/-25% band by 9.58 pp. baseline 547.36 ns, observed 358.06 ns.
    • [Verbara.Sdk.Benchmarks.ChannelManagerBenchmark.Create1000Channels] FASTER than baseline by 30.11%, outside the +/-30% band by 0.11 pp. baseline 353056 ns, observed 246751.16 ns.
    • [Verbara.Sdk.Benchmarks.ChannelManagerBenchmark.LookupByName] FASTER than baseline by 34.77%, outside the +/-15% band by 19.77 pp. baseline 14.69 ns, observed 9.58 ns.
    • [Verbara.Sdk.Benchmarks.ChannelManagerBenchmark.LookupByUniqueId] FASTER than baseline by 33.75%, outside the +/-15% band by 18.75 pp. baseline 12.01 ns, observed 7.96 ns.
    • [Verbara.Sdk.Benchmarks.ChannelManagerBenchmark.Update1000ChannelStates] FASTER than baseline by 16.46%, outside the +/-15% band by 1.46 pp. baseline 65925 ns, observed 55074.90 ns.
    • [Verbara.Sdk.Benchmarks.ObserverDispatchBenchmark.Dispatch_100Observers] FASTER than baseline by 16.80%, outside the +/-15% band by 1.80 pp. baseline 47.61 ns, observed 39.61 ns.
    • [Verbara.Sdk.Benchmarks.ObserverDispatchBenchmark.Dispatch_10Observers] FASTER than baseline by 30.06%, outside the +/-20% band by 10.06 pp. baseline 5.67 ns, observed 3.97 ns.
    • [Verbara.Sdk.Benchmarks.ObserverDispatchBenchmark.Dispatch_1Observer] FASTER than baseline by 77.19%, outside the +/-15% band by 62.19 pp. baseline 0.94 ns, observed 0.21 ns.

    Measured

    benchmark baseline (ns) observed (ns) delta band
    AmiProtocolReaderBenchmark.Parse100EventBatch 88642 55801.18 -37.05% ±10% BREACH (FASTER)
    AmiProtocolReaderBenchmark.ParseResponse 981.16 619.83 -36.83% ±25% BREACH (FASTER)
    AmiProtocolReaderBenchmark.ParseSingleEvent 1484 907.93 -38.82% ±25% BREACH (FASTER)
    AriJsonBenchmark.Deserialize100Channels 64354 37622.70 -41.54% ±15% BREACH (FASTER)
    AriJsonBenchmark.DeserializeBridge 562.22 418.75 -25.52% ±30% ok
    AriJsonBenchmark.DeserializeChannel 666.79 388.01 -41.81% ±20% BREACH (FASTER)
    AriJsonBenchmark.SerializeBridge 264.5 175.12 -33.79% ±35% ok
    AriJsonBenchmark.SerializeChannel 363.01 232.44 -35.97% ±45% ok
    AriParseEventBenchmark.ParseChannelDtmf 1939 1272.63 -34.37% ±25% BREACH (FASTER)
    AriParseEventBenchmark.ParseStasisStart 3947 2497.34 -36.73% ±25% BREACH (FASTER)
    AriParseEventBenchmark.ParseUnknownEvent 547.36 358.06 -34.58% ±25% BREACH (FASTER)
    ChannelManagerBenchmark.Create1000Channels 353056 246751.16 -30.11% ±30% BREACH (FASTER)
    ChannelManagerBenchmark.EnumerateByState 125710 100104.33 -20.37% ±30% ok
    ChannelManagerBenchmark.LookupByName 14.69 9.58 -34.77% ±15% BREACH (FASTER)
    ChannelManagerBenchmark.LookupByUniqueId 12.01 7.96 -33.75% ±15% BREACH (FASTER)
    ChannelManagerBenchmark.Update1000ChannelStates 65925 55074.90 -16.46% ±15% BREACH (FASTER)
    ObserverDispatchBenchmark.Dispatch_100Observers 47.61 39.61 -16.80% ±15% BREACH (FASTER)
    ObserverDispatchBenchmark.Dispatch_10Observers 5.67 3.97 -30.06% ±20% BREACH (FASTER)
    ObserverDispatchBenchmark.Dispatch_1Observer 0.94 0.21 -77.19% ±15% BREACH (FASTER)

    Measured but not baselined (4)

    Not a failure. Add a benchmarks entry once its hosted-runner spread is observed — see baseline.README.md.

    • Verbara.Sdk.Benchmarks.TurnDetectionBenchmark.EndOfUtteranceDecision(UtteranceSeconds: 1)
    • Verbara.Sdk.Benchmarks.TurnDetectionBenchmark.EndOfUtteranceDecision(UtteranceSeconds: 2)
    • Verbara.Sdk.Benchmarks.TurnDetectionBenchmark.EndOfUtteranceDecision(UtteranceSeconds: 4)
    • Verbara.Sdk.Benchmarks.TurnDetectionBenchmark.EndOfUtteranceDecision(UtteranceSeconds: 8)

    Run: https://github.com/verbara/Verbara.Sdk/actions/runs/35536542837

    Baselines move only by human-authored commit (ADR-0042 D6) — CI never writes back to
    Tests/Verbara.Sdk.Benchmarks/baseline.json. If this is a real regression, fix the code.
    If the hosted-runner fleet moved under us, re-baseline per
    Tests/Verbara.Sdk.Benchmarks/baseline.README.md: a PR stating which benchmark moved, in
    which direction, by how much, and why.

  2. github-actions commented on Sep 20, 2026

    @github-actions
    github-actionsbot
    on Sep 20, 2026 – with GitHub Actions
    Author

    Perf regression vs Tests/Verbara.Sdk.Benchmarks/baseline.json

    Mode: OBSERVING (reports, does not fail the job — §2.5)

    Band breaches (5)

    • [Verbara.Sdk.Benchmarks.AmiProtocolReaderBenchmark.Parse100EventBatch] FASTER than baseline by 21.46%, outside the +/-10% band by 11.46 pp. baseline 88642 ns, observed 69621.17 ns.
    • [Verbara.Sdk.Benchmarks.AriJsonBenchmark.Deserialize100Channels] FASTER than baseline by 24.39%, outside the +/-15% band by 9.39 pp. baseline 64354 ns, observed 48660.41 ns.
    • [Verbara.Sdk.Benchmarks.AriJsonBenchmark.DeserializeChannel] FASTER than baseline by 25.27%, outside the +/-20% band by 5.27 pp. baseline 666.79 ns, observed 498.28 ns.
    • [Verbara.Sdk.Benchmarks.ObserverDispatchBenchmark.Dispatch_100Observers] FASTER than baseline by 19.88%, outside the +/-15% band by 4.88 pp. baseline 47.61 ns, observed 38.14 ns.
    • [Verbara.Sdk.Benchmarks.ObserverDispatchBenchmark.Dispatch_1Observer] FASTER than baseline by 66.49%, outside the +/-15% band by 51.49 pp. baseline 0.94 ns, observed 0.31 ns.

    Measured

    benchmark baseline (ns) observed (ns) delta band
    AmiProtocolReaderBenchmark.Parse100EventBatch 88642 69621.17 -21.46% ±10% BREACH (FASTER)
    AmiProtocolReaderBenchmark.ParseResponse 981.16 848.67 -13.50% ±25% ok
    AmiProtocolReaderBenchmark.ParseSingleEvent 1484 1153.76 -22.25% ±25% ok
    AriJsonBenchmark.Deserialize100Channels 64354 48660.41 -24.39% ±15% BREACH (FASTER)
    AriJsonBenchmark.DeserializeBridge 562.22 474.49 -15.60% ±30% ok
    AriJsonBenchmark.DeserializeChannel 666.79 498.28 -25.27% ±20% BREACH (FASTER)
    AriJsonBenchmark.SerializeBridge 264.5 223.81 -15.38% ±35% ok
    AriJsonBenchmark.SerializeChannel 363.01 290.19 -20.06% ±45% ok
    AriParseEventBenchmark.ParseChannelDtmf 1939 1694.65 -12.60% ±25% ok
    AriParseEventBenchmark.ParseStasisStart 3947 3341.81 -15.33% ±25% ok
    AriParseEventBenchmark.ParseUnknownEvent 547.36 481.52 -12.03% ±25% ok
    ChannelManagerBenchmark.Create1000Channels 353056 302697.20 -14.26% ±30% ok
    ChannelManagerBenchmark.EnumerateByState 125710 123901.36 -1.44% ±30% ok
    ChannelManagerBenchmark.LookupByName 14.69 15.03 +2.32% ±15% ok
    ChannelManagerBenchmark.LookupByUniqueId 12.01 10.35 -13.84% ±15% ok
    ChannelManagerBenchmark.Update1000ChannelStates 65925 69736.61 +5.78% ±15% ok
    ObserverDispatchBenchmark.Dispatch_100Observers 47.61 38.14 -19.88% ±15% BREACH (FASTER)
    ObserverDispatchBenchmark.Dispatch_10Observers 5.67 4.76 -16.09% ±20% ok
    ObserverDispatchBenchmark.Dispatch_1Observer 0.94 0.31 -66.49% ±15% BREACH (FASTER)

    Measured but not baselined (4)

    Not a failure. Add a benchmarks entry once its hosted-runner spread is observed — see baseline.README.md.

    • Verbara.Sdk.Benchmarks.TurnDetectionBenchmark.EndOfUtteranceDecision(UtteranceSeconds: 1)
    • Verbara.Sdk.Benchmarks.TurnDetectionBenchmark.EndOfUtteranceDecision(UtteranceSeconds: 2)
    • Verbara.Sdk.Benchmarks.TurnDetectionBenchmark.EndOfUtteranceDecision(UtteranceSeconds: 4)
    • Verbara.Sdk.Benchmarks.TurnDetectionBenchmark.EndOfUtteranceDecision(UtteranceSeconds: 8)

    Run: https://github.com/verbara/Verbara.Sdk/actions/runs/35536628224

    Baselines move only by human-authored commit (ADR-0042 D6) — CI never writes back to
    Tests/Verbara.Sdk.Benchmarks/baseline.json. If this is a real regression, fix the code.
    If the hosted-runner fleet moved under us, re-baseline per
    Tests/Verbara.Sdk.Benchmarks/baseline.README.md: a PR stating which benchmark moved, in
    which direction, by how much, and why.

  3. github-actions commented on Oct 4, 2026

    @github-actions
    github-actionsbot
    on Oct 4, 2026 – with GitHub Actions
    Author

    Perf regression vs Tests/Verbara.Sdk.Benchmarks/baseline.json

    Mode: OBSERVING (reports, does not fail the job — §2.5)

    Band breaches (18)

    • [Verbara.Sdk.Benchmarks.AmiProtocolReaderBenchmark.Parse100EventBatch] FASTER than baseline by 48.70%, outside the +/-10% band by 38.70 pp. baseline 88642 ns, observed 45477.00 ns.
    • [Verbara.Sdk.Benchmarks.AmiProtocolReaderBenchmark.ParseResponse] FASTER than baseline by 47.30%, outside the +/-25% band by 22.30 pp. baseline 981.16 ns, observed 517.08 ns.
    • [Verbara.Sdk.Benchmarks.AmiProtocolReaderBenchmark.ParseSingleEvent] FASTER than baseline by 49.58%, outside the +/-25% band by 24.58 pp. baseline 1484 ns, observed 748.19 ns.
    • [Verbara.Sdk.Benchmarks.AriJsonBenchmark.Deserialize100Channels] FASTER than baseline by 47.01%, outside the +/-15% band by 32.01 pp. baseline 64354 ns, observed 34099.30 ns.
    • [Verbara.Sdk.Benchmarks.AriJsonBenchmark.DeserializeBridge] FASTER than baseline by 41.78%, outside the +/-30% band by 11.78 pp. baseline 562.22 ns, observed 327.33 ns.
    • [Verbara.Sdk.Benchmarks.AriJsonBenchmark.DeserializeChannel] FASTER than baseline by 48.84%, outside the +/-20% band by 28.84 pp. baseline 666.79 ns, observed 341.14 ns.
    • [Verbara.Sdk.Benchmarks.AriJsonBenchmark.SerializeBridge] FASTER than baseline by 43.61%, outside the +/-35% band by 8.61 pp. baseline 264.5 ns, observed 149.16 ns.
    • [Verbara.Sdk.Benchmarks.AriJsonBenchmark.SerializeChannel] FASTER than baseline by 47.91%, outside the +/-45% band by 2.91 pp. baseline 363.01 ns, observed 189.10 ns.
    • [Verbara.Sdk.Benchmarks.AriParseEventBenchmark.ParseChannelDtmf] FASTER than baseline by 43.27%, outside the +/-25% band by 18.27 pp. baseline 1939 ns, observed 1099.96 ns.
    • [Verbara.Sdk.Benchmarks.AriParseEventBenchmark.ParseStasisStart] FASTER than baseline by 47.01%, outside the +/-25% band by 22.01 pp. baseline 3947 ns, observed 2091.35 ns.
    • [Verbara.Sdk.Benchmarks.AriParseEventBenchmark.ParseUnknownEvent] FASTER than baseline by 49.42%, outside the +/-25% band by 24.42 pp. baseline 547.36 ns, observed 276.86 ns.
    • [Verbara.Sdk.Benchmarks.ChannelManagerBenchmark.Create1000Channels] FASTER than baseline by 33.30%, outside the +/-30% band by 3.30 pp. baseline 353056 ns, observed 235482.91 ns.
    • [Verbara.Sdk.Benchmarks.ChannelManagerBenchmark.EnumerateByState] FASTER than baseline by 46.34%, outside the +/-30% band by 16.34 pp. baseline 125710 ns, observed 67460.15 ns.
    • [Verbara.Sdk.Benchmarks.ChannelManagerBenchmark.LookupByName] FASTER than baseline by 39.93%, outside the +/-15% band by 24.93 pp. baseline 14.69 ns, observed 8.82 ns.
    • [Verbara.Sdk.Benchmarks.ChannelManagerBenchmark.LookupByUniqueId] FASTER than baseline by 50.80%, outside the +/-15% band by 35.80 pp. baseline 12.01 ns, observed 5.91 ns.
    • [Verbara.Sdk.Benchmarks.ChannelManagerBenchmark.Update1000ChannelStates] FASTER than baseline by 36.07%, outside the +/-15% band by 21.07 pp. baseline 65925 ns, observed 42148.71 ns.
    • [Verbara.Sdk.Benchmarks.ObserverDispatchBenchmark.Dispatch_100Observers] FASTER than baseline by 39.78%, outside the +/-15% band by 24.78 pp. baseline 47.61 ns, observed 28.67 ns.
    • [Verbara.Sdk.Benchmarks.ObserverDispatchBenchmark.Dispatch_10Observers] FASTER than baseline by 59.97%, outside the +/-20% band by 39.97 pp. baseline 5.67 ns, observed 2.27 ns.

    Measured

    benchmark baseline (ns) observed (ns) delta band
    AmiProtocolReaderBenchmark.Parse100EventBatch 88642 45477.00 -48.70% ±10% BREACH (FASTER)
    AmiProtocolReaderBenchmark.ParseResponse 981.16 517.08 -47.30% ±25% BREACH (FASTER)
    AmiProtocolReaderBenchmark.ParseSingleEvent 1484 748.19 -49.58% ±25% BREACH (FASTER)
    AriJsonBenchmark.Deserialize100Channels 64354 34099.30 -47.01% ±15% BREACH (FASTER)
    AriJsonBenchmark.DeserializeBridge 562.22 327.33 -41.78% ±30% BREACH (FASTER)
    AriJsonBenchmark.DeserializeChannel 666.79 341.14 -48.84% ±20% BREACH (FASTER)
    AriJsonBenchmark.SerializeBridge 264.5 149.16 -43.61% ±35% BREACH (FASTER)
    AriJsonBenchmark.SerializeChannel 363.01 189.10 -47.91% ±45% BREACH (FASTER)
    AriParseEventBenchmark.ParseChannelDtmf 1939 1099.96 -43.27% ±25% BREACH (FASTER)
    AriParseEventBenchmark.ParseStasisStart 3947 2091.35 -47.01% ±25% BREACH (FASTER)
    AriParseEventBenchmark.ParseUnknownEvent 547.36 276.86 -49.42% ±25% BREACH (FASTER)
    ChannelManagerBenchmark.Create1000Channels 353056 235482.91 -33.30% ±30% BREACH (FASTER)
    ChannelManagerBenchmark.EnumerateByState 125710 67460.15 -46.34% ±30% BREACH (FASTER)
    ChannelManagerBenchmark.LookupByName 14.69 8.82 -39.93% ±15% BREACH (FASTER)
    ChannelManagerBenchmark.LookupByUniqueId 12.01 5.91 -50.80% ±15% BREACH (FASTER)
    ChannelManagerBenchmark.Update1000ChannelStates 65925 42148.71 -36.07% ±15% BREACH (FASTER)
    ObserverDispatchBenchmark.Dispatch_100Observers 47.61 28.67 -39.78% ±15% BREACH (FASTER)
    ObserverDispatchBenchmark.Dispatch_10Observers 5.67 2.27 -59.97% ±20% BREACH (FASTER)

    Measured but not baselined (5)

    Not a failure. Add a benchmarks entry once its hosted-runner spread is observed — see baseline.README.md.

    • Verbara.Sdk.Benchmarks.ObserverDispatchBenchmark.Dispatch_1Observer
    • Verbara.Sdk.Benchmarks.TurnDetectionBenchmark.EndOfUtteranceDecision(UtteranceSeconds: 1)
    • Verbara.Sdk.Benchmarks.TurnDetectionBenchmark.EndOfUtteranceDecision(UtteranceSeconds: 2)
    • Verbara.Sdk.Benchmarks.TurnDetectionBenchmark.EndOfUtteranceDecision(UtteranceSeconds: 4)
    • Verbara.Sdk.Benchmarks.TurnDetectionBenchmark.EndOfUtteranceDecision(UtteranceSeconds: 8)

    Run: https://github.com/verbara/Verbara.Sdk/actions/runs/37195447499

    Baselines move only by human-authored commit (ADR-0042 D6) — CI never writes back to
    Tests/Verbara.Sdk.Benchmarks/baseline.json. If this is a real regression, fix the code.
    If the hosted-runner fleet moved under us, re-baseline per
    Tests/Verbara.Sdk.Benchmarks/baseline.README.md: a PR stating which benchmark moved, in
    which direction, by how much, and why.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions