Performance reference results
Dense chart and CSV warehouse for a single reference host (maintainer-ws-1).
For tuning decisions, start from
Performance findings and
Tuning evidence (studies); use this page when you
need every published figure in one place. Prefer reading each chart relative to
its baseline cells on that host. These are not service-level objectives.
Reproduce on your hardware with the
harness instructions before making capacity decisions.
Absolute QPS is not a portable cross-host capacity claim.
Reading forward_slow cells — saturation against a 50 ms backend
forward_slow cells offer far more
concurrency than a runtime that blocks on upstream latency can absorb, which is
the point: they show what happens to each runtime model when the backend is slow
and the client keeps asking. Read both columns together. A runtime that
multiplexes in-flight queries answers near the upstream delay itself; a runtime
that occupies a worker for the whole round trip reports a small fraction of that
throughput and an average latency of seconds, because queries wait in Conduit
rather than at the backend. Every published cell here is measured on a stub
upstream that stays well clear of saturation and is checked for successful
answers, so the numbers are Conduit's behavior, not the harness reaching its
limit. See
How load is applied
and Only successful answers count.
Generated 2026-08-14T22:39:28Z from committed reference JSON (no live load suite in docs CI).
Lab profile
| Field | Value |
|---|---|
| Profile id | maintainer-ws-1 |
| Display name | Maintainer workstation (maintainer-ws-1) |
| CPU | Intel(R) Core(TM) i9-14900K |
| Cores (physical / logical) | 32 / 32 |
| OS | Linux 6.8.0-136-generic |
| Conduit | target/release/conduit (unknown) |
| Loadgen | dnsperf mode=docker |
| Run generated_at | 2026-08-10T17:39:26Z |
Underlying JSON: perf/results/references/ (see latest-reference.json pointer in a checkout).
Decision context: Performance findings · Tuning evidence (studies). Scenario glosses: Performance scenarios (glossary, not a browse-first surface).
Scale
Achieved QPS — sync vs split_io (forward_fast)
| Scenario | Runtime | Achieved QPS | Avg latency (ms) | Sent | Completed | Lost | Workers |
|---|---|---|---|---|---|---|---|
| scale-sync-forward-fast | sync | 76269.9 | 26.1 | 765379 | 765379 | 0 | ingress=2 |
| scale-split-io-forward-fast | split_io | 138744.2 | 14.4 | 1389252 | 1389252 | 0 | ingress=2, policy=2, io=2 |
Achieved QPS — sync vs split_io (forward_slow)
| Scenario | Runtime | Achieved QPS | Avg latency (ms) | Sent | Completed | Lost | Workers |
|---|---|---|---|---|---|---|---|
| scale-sync-forward-slow | sync | 5.7 | 2508.7 | 12188 | 198 | 11990 | ingress=2 |
| scale-split-io-forward-slow | split_io | 39080.2 | 51.1 | 1174342 | 1174342 | 0 | ingress=2, policy=2, io=2 |
Achieved QPS — sync cache_hit
| Scenario | Runtime | Achieved QPS | Avg latency (ms) | Sent | Completed | Lost | Workers |
|---|---|---|---|---|---|---|---|
| scale-sync-cache-hit | sync | 329202.8 | 6.1 | 3294253 | 3294253 | 0 | ingress=2 |
Achieved QPS — split_io topology (forward_fast)
| Scenario | Runtime | Achieved QPS | Avg latency (ms) | Sent | Completed | Lost | Workers |
|---|---|---|---|---|---|---|---|
| scale-split-io-forward-fast | split_io | 138744.2 | 14.4 | 1389252 | 1389252 | 0 | ingress=2, policy=2, io=2 |
| scale-split-io-topology-heavy | split_io | 258908.7 | 7.7 | 2590870 | 2590870 | 0 | ingress=4, policy=4, io=4 |
Shutdown drain
Drain duration under forward_slow
| Drain policy | Drain duration (ms) | Client failures during stop | QPS | Avg latency (ms) | Sent | Completed |
|---|---|---|---|---|---|---|
| drain_complete | 113.5 | 299 | 6.5 | 996.6 | 377 | 78 |
| drain_budgeted | 113.6 | 200 | 10.0 | 909.2 | 280 | 80 |
| drain_minimal | 63.7 | 200 | 9.7 | 902.1 | 278 | 78 |
Feature tax
Feature tax — metrics scrape tax (forward_fast)
| Posture | Achieved QPS | Avg latency (ms) | Sent | Completed | Lost |
|---|---|---|---|---|---|
| metrics_off | 76124.1 | 26.2 | 763843 | 763843 | 0 |
| minimal_scrape | 73038.1 | 27.3 | 732337 | 732337 | 0 |
| standard_scrape | 70157.8 | 28.4 | 703550 | 703550 | 0 |
Feature tax — dnstap off / sampled / full (forward_fast)
| Posture | Achieved QPS | Avg latency (ms) | Sent | Completed | Lost |
|---|---|---|---|---|---|
| dnstap_off | 77696.7 | 25.7 | 778975 | 778975 | 0 |
| dnstap_sampled | 73759.2 | 27.1 | 739578 | 739578 | 0 |
| dnstap_full | 68939.9 | 28.9 | 692386 | 692386 | 0 |
Feature tax — collect vs emit (forward_fast)
| Posture | Achieved QPS | Avg latency (ms) | Sent | Completed | Lost |
|---|---|---|---|---|---|
| no_collect | 73615.4 | 27.1 | 738135 | 738135 | 0 |
| collect_only | 70398.5 | 28.3 | 706795 | 706795 | 0 |
| collect_emit | 70635.6 | 28.2 | 708920 | 708920 | 0 |
Feature tax — metrics and dnstap combined (forward_fast)
| Posture | Achieved QPS | Avg latency (ms) | Sent | Completed | Lost |
|---|---|---|---|---|---|
| metrics_off | 75940.7 | 26.3 | 761386 | 761386 | 0 |
| standard_scrape | 70029.9 | 28.5 | 703153 | 703153 | 0 |
| dnstap_full | 68939.9 | 28.9 | 692386 | 692386 | 0 |
| standard_dnstap_full | 64339.4 | 31.0 | 645499 | 645499 | 0 |
Feature tax — scrape hammer under load (forward_fast)
| Posture | Achieved QPS | Avg latency (ms) | Sent | Completed | Lost |
|---|---|---|---|---|---|
| metrics_off | 75940.7 | 26.3 | 761386 | 761386 | 0 |
| standard_scrape | 70029.9 | 28.5 | 703153 | 703153 | 0 |
| scrape_hammer | 57968.8 | 34.4 | 582917 | 582917 | 0 |
Feature tax — metrics scrape under split_io (forward_fast)
| Posture | Achieved QPS | Avg latency (ms) | Sent | Completed | Lost |
|---|---|---|---|---|---|
| metrics_off | 146802.2 | 13.6 | 1469986 | 1469986 | 0 |
| standard_scrape | 139823.3 | 14.3 | 1400096 | 1400096 | 0 |