Stress Test Results

Long-running stress tests comparing sustained performance between Dekaf and Confluent.Kafka under real-world load.

Last Updated: 2026-08-05 02:11 UTC

info

The paired Dekaf vs Confluent comparison runs weekly (Sunday 2 AM UTC) and updates this page. Manual dispatches stay Dekaf-only unless full_run explicitly requests the same paired publish path. Tests measure sustained performance over 15+ minutes with real Kafka instances.

At a glance

Each row is a like-for-like comparison: both clients run the same sustained workload sequentially on the same VM, and repeated samples are aggregated with a geometric mean across both run orders.

Scenario	Dekaf	Confluent	Throughput	CPU per message
Produce — fire-and-forget	1,589,592 msg/s	1,364,593 msg/s	1.2× faster	1.4× less
Produce — fire-and-forget (3 brokers)	1,202,782 msg/s	896,773 msg/s	1.3× faster	1.6× less
Produce — acks=all	1,588,223 msg/s	1,401,374 msg/s	1.1× faster	1.4× less
Produce — acks=all (3 brokers)	852,910 msg/s	769,005 msg/s	1.1× faster	1.3× less
Produce — fire-and-forget, idempotent	1,617,459 msg/s	1,450,111 msg/s	1.1× faster	1.4× less
Produce — fire-and-forget, idempotent (3 brokers)	1,117,158 msg/s	858,416 msg/s	1.3× faster	1.6× less
Produce + consume round-trip	2,652,540 msg/s	1,731,781 msg/s	1.5× faster	2.0× less
Produce — transactional (exactly-once) (3 brokers)	1,244 msg/s	174 msg/s	7.1× faster	1.1× less
Consume — messages	1,771,526 msg/s	1,365,730 msg/s	1.3× faster	1.4× less
Consume — batches	1,195,583 msg/s	—	—	—
Consume — raw bytes	3,647,900 msg/s	—	—	—
Consume — raw byte batches	4,178,857 msg/s	—	—	—

"On par" means within ±5% — differences that small are run-to-run noise. "CPU per message" compares the client CPU cost of delivering one message; "less" means Dekaf needs less CPU. Rows showing "—" have no Confluent counterpart in this run (for example, batch and raw consume APIs that librdkafka does not expose). The full per-run data is below.

Full results

Each section holds the measured per-run data behind the summary: repeated same-VM samples in both client orders, CPU per message and per request, and throughput drift across the run.

Producer (Fire-and-Forget) (15 minutes, 1000B messages)

Order-Balanced Aggregate

Client	Samples	Geomean comparison msg/s	Sample range	Median CPU μs/msg	Comparison Ratio
Dekaf	2	1,589,592	1,567,598–1,611,895	0.94	1.16x
Confluent	2	1,364,593	1,334,107–1,395,777	1.34	1.00x

The aggregate uses the geometric mean across balanced same-VM samples run in both dekaf-first and confluent-first order. Raw ordered samples remain below.

Client	CPU μs/msg	CPU μs/request	Messages/sec	Median msg/s	Drift	Slope %/min	MB/sec	Accepted msg/s	Errors	Standing cores
Dekaf (3conn)	0.71	688.85	1,916,810	1,971,393	+15.7%	+1.55%	1828.01	1,916,810	0	1.36
Dekaf (dekaf-first)	0.95	974.98	1,589,098	1,611,895	+1.5%	+0.07%	1515.48	1,589,098	0	1.51
Dekaf (confluent-first)	0.92	943.22	1,541,970	1,567,598	-6.3%	-0.50%	1470.54	1,541,970	0	1.42
Confluent (dekaf-first)	1.28	-	1,384,814	1,395,777	-0.5%	-0.07%	1320.66	1,384,814	0	1.77
Confluent (confluent-first)	1.40	-	1,323,202	1,334,107	+1.3%	+0.12%	1261.90	1,323,202	0	1.85

Median msg/s is the median sampled client-side throughput interval; it shows steady-state throughput without letting a short late-run stall dominate the whole-run average.

Rows and Comparison Ratio use Median msg/s when available; older result files without interval samples fall back to Messages/sec.

Drift compares last-third with first-third average throughput. Slope is the normalized least-squares trend; steady-state below 85% of peak or slope below -1%/min fails the regression gate.

Messages/sec counts broker-confirmed deliveries (end-offset delta). Accepted msg/s is the client-side append rate — a large gap means messages were buffered or dropped without ever reaching the broker.

Producer (Fire-and-Forget), 3 Brokers (15 minutes, 1000B messages)

Client	CPU μs/msg	CPU μs/request	Messages/sec	Median msg/s	Drift	Slope %/min	MB/sec	Accepted msg/s	Errors	Standing cores
Dekaf (3conn)	1.03	992.58	1,251,395	1,250,962	+2.6%	+0.26%	1193.42	1,251,395	0	1.29
Dekaf	1.05	1010.85	1,198,260	1,202,782	+4.3%	+0.43%	1142.75	1,198,260	0	1.26
Confluent	1.68	-	895,991	896,773	+0.3%	+0.02%	854.48	895,991	0	1.51

Median msg/s is the median sampled client-side throughput interval; it shows steady-state throughput without letting a short late-run stall dominate the whole-run average.

Rows and Comparison Ratio use Median msg/s when available; older result files without interval samples fall back to Messages/sec.

Drift compares last-third with first-third average throughput. Slope is the normalized least-squares trend; steady-state below 85% of peak or slope below -1%/min fails the regression gate.

Messages/sec counts broker-confirmed deliveries (end-offset delta). Accepted msg/s is the client-side append rate — a large gap means messages were buffered or dropped without ever reaching the broker.

Producer (Acks All) (15 minutes, 1000B messages)

Order-Balanced Aggregate

Client	Samples	Geomean comparison msg/s	Sample range	Median CPU μs/msg	Comparison Ratio
Dekaf	2	1,588,223	1,572,835–1,603,762	0.94	1.13x
Confluent	2	1,401,374	1,384,156–1,418,807	1.27	1.00x

The aggregate uses the geometric mean across balanced same-VM samples run in both dekaf-first and confluent-first order. Raw ordered samples remain below.

Client	CPU μs/msg	CPU μs/request	Messages/sec	Median msg/s	Drift	Slope %/min	MB/sec	Accepted msg/s	Errors	Standing cores
Dekaf (dekaf-first)	0.94	966.90	1,587,185	1,603,762	+1.5%	+0.11%	1513.66	1,587,185	0	1.49
Dekaf (confluent-first)	0.93	956.65	1,560,040	1,572,835	-0.6%	-0.03%	1487.77	1,560,040	0	1.46
Confluent (confluent-first)	1.25	-	1,413,322	1,418,807	+0.9%	+0.09%	1347.85	1,413,322	0	1.77
Confluent (dekaf-first)	1.29	-	1,375,551	1,384,156	-2.9%	-0.25%	1311.83	1,375,551	0	1.78

Median msg/s is the median sampled client-side throughput interval; it shows steady-state throughput without letting a short late-run stall dominate the whole-run average.

Rows and Comparison Ratio use Median msg/s when available; older result files without interval samples fall back to Messages/sec.

Drift compares last-third with first-third average throughput. Slope is the normalized least-squares trend; steady-state below 85% of peak or slope below -1%/min fails the regression gate.

Messages/sec counts broker-confirmed deliveries (end-offset delta). Accepted msg/s is the client-side append rate — a large gap means messages were buffered or dropped without ever reaching the broker.

Producer (Acks All), 3 Brokers (15 minutes, 1000B messages)

Client	CPU μs/msg	CPU μs/request	Messages/sec	Median msg/s	Drift	Slope %/min	MB/sec	Accepted msg/s	Errors	Standing cores
Dekaf	1.50	1499.04	849,533	852,910	+5.2%	+0.54%	810.18	849,533	0	1.28
Confluent	1.96	-	772,783	769,005	+0.3%	+0.07%	736.98	772,783	0	1.52

Median msg/s is the median sampled client-side throughput interval; it shows steady-state throughput without letting a short late-run stall dominate the whole-run average.

Rows and Comparison Ratio use Median msg/s when available; older result files without interval samples fall back to Messages/sec.

Drift compares last-third with first-third average throughput. Slope is the normalized least-squares trend; steady-state below 85% of peak or slope below -1%/min fails the regression gate.

Messages/sec counts broker-confirmed deliveries (end-offset delta). Accepted msg/s is the client-side append rate — a large gap means messages were buffered or dropped without ever reaching the broker.

Producer (Fire-and-Forget, Idempotent) (15 minutes, 1000B messages)

Order-Balanced Aggregate

Client	Samples	Geomean comparison msg/s	Sample range	Median CPU μs/msg	Comparison Ratio
Dekaf	2	1,617,459	1,570,640–1,665,674	0.91	1.12x
Confluent	2	1,450,111	1,441,199–1,459,079	1.25	1.00x

The aggregate uses the geometric mean across balanced same-VM samples run in both dekaf-first and confluent-first order. Raw ordered samples remain below.

Client	CPU μs/msg	CPU μs/request	Messages/sec	Median msg/s	Drift	Slope %/min	MB/sec	Accepted msg/s	Errors	Standing cores
Dekaf (dekaf-first)	0.89	916.60	1,641,390	1,665,674	+1.4%	+0.11%	1565.35	1,641,390	0	1.47
Dekaf (3conn)	0.77	711.85	1,550,532	1,577,926	+10.6%	+0.90%	1478.70	1,550,532	0	1.19
Dekaf (confluent-first)	0.93	952.77	1,543,502	1,570,640	-6.4%	-0.51%	1472.00	1,543,502	0	1.44
Confluent (dekaf-first)	1.25	-	1,444,531	1,459,079	+1.4%	+0.06%	1377.61	1,444,531	0	1.80
Confluent (confluent-first)	1.26	-	1,411,305	1,441,199	+2.0%	-0.13%	1345.93	1,411,305	0	1.78

Median msg/s is the median sampled client-side throughput interval; it shows steady-state throughput without letting a short late-run stall dominate the whole-run average.

Rows and Comparison Ratio use Median msg/s when available; older result files without interval samples fall back to Messages/sec.

Drift compares last-third with first-third average throughput. Slope is the normalized least-squares trend; steady-state below 85% of peak or slope below -1%/min fails the regression gate.

Messages/sec counts broker-confirmed deliveries (end-offset delta). Accepted msg/s is the client-side append rate — a large gap means messages were buffered or dropped without ever reaching the broker.

Producer (Fire-and-Forget, Idempotent), 3 Brokers (15 minutes, 1000B messages)

Client	CPU μs/msg	CPU μs/request	Messages/sec	Median msg/s	Drift	Slope %/min	MB/sec	Accepted msg/s	Errors	Standing cores
Dekaf (3conn)	1.09	1095.57	1,199,638	1,208,997	+0.8%	+0.09%	1144.06	1,199,638	0	1.30
Dekaf	1.17	1181.56	1,105,422	1,117,158	+4.2%	+0.43%	1054.21	1,105,422	0	1.29
Confluent	1.82	-	858,561	858,416	-0.0%	+0.00%	818.79	858,561	0	1.57

Median msg/s is the median sampled client-side throughput interval; it shows steady-state throughput without letting a short late-run stall dominate the whole-run average.

Rows and Comparison Ratio use Median msg/s when available; older result files without interval samples fall back to Messages/sec.

Drift compares last-third with first-third average throughput. Slope is the normalized least-squares trend; steady-state below 85% of peak or slope below -1%/min fails the regression gate.

Messages/sec counts broker-confirmed deliveries (end-offset delta). Accepted msg/s is the client-side append rate — a large gap means messages were buffered or dropped without ever reaching the broker.

Producer → Consumer Round-Trip Steady State (15 minutes, 128B messages)

Client	CPU μs/msg	CPU μs/request	Messages/sec	Median msg/s	Drift	Slope %/min	MB/sec	Accepted msg/s	Errors	Standing cores
Dekaf	0.93	5003.84	1,399,723	2,652,540	+45.8%	+459.80%	170.86	1,399,723	0	1.30
Confluent	1.85	-	125,543	1,731,781	+28.5%	+233.66%	15.33	125,543	0	0.23

Median msg/s is the median sampled client-side throughput interval; it shows steady-state throughput without letting a short late-run stall dominate the whole-run average.

Rows and Comparison Ratio use Median msg/s when available; older result files without interval samples fall back to Messages/sec.

Drift compares last-third with first-third average throughput. Slope is the normalized least-squares trend; steady-state below 85% of peak or slope below -1%/min fails the regression gate.

Messages/sec counts broker-confirmed deliveries (end-offset delta). Accepted msg/s is the client-side append rate — a large gap means messages were buffered or dropped without ever reaching the broker.

Round-Trip Validation

Client	Expected	Consumed	Missing	Duplicates	Corrupt	Out of Order	Wrong Partition	Unexpected	Timed Out	Result
Confluent	19,792,477	19,792,477	0	0	0	0	0	0	no	PASS
Dekaf	19,792,477	19,792,477	0	0	0	0	0	0	no	PASS

Producer (Transactional EOS), 3 Brokers (15 minutes, 1000B messages)

Client	CPU μs/msg	CPU μs/request	Messages/sec	Median msg/s	Drift	Slope %/min	MB/sec	Accepted msg/s	Errors	Standing cores
Dekaf	205.90	205.90	925	1,244	+10.3%	+0.89%	0.88	1,234	0	0.25
Confluent	231.28	-	131	174	+0.2%	+0.03%	0.12	175	0	0.04

Median msg/s is the median sampled client-side throughput interval; it shows steady-state throughput without letting a short late-run stall dominate the whole-run average.

Rows and Comparison Ratio use Median msg/s when available; older result files without interval samples fall back to Messages/sec.

Drift compares last-third with first-third average throughput. Slope is the normalized least-squares trend; steady-state below 85% of peak or slope below -1%/min fails the regression gate.

Messages/sec counts broker-confirmed deliveries (end-offset delta). Accepted msg/s is the client-side append rate — a large gap means messages were buffered or dropped without ever reaching the broker.

Transaction Verification

Client	Accepted	Committed	Aborted	Delivered	Duplicates	Shortfall	Aborted leaks	Unexpected	Missing sentinels	Status
Confluent	157,200	117,900	39,300	117,900	0	0	0	0	0	PASS
Dekaf	1,110,500	832,900	277,600	832,900	0	0	0	0	0	PASS

Consumer (15 minutes, 1000B messages, 16,384B seed batches)

Client	CPU μs/msg	CPU μs/request	Messages/sec	Median msg/s	Drift	Slope %/min	MB/sec	Accepted msg/s	Errors	Standing cores
Dekaf	0.74	-	1,769,310	1,771,526	-1.6%	-0.20%	1687.35	-	0	1.31
Confluent	1.07	-	1,359,428	1,365,730	+1.6%	+0.17%	1296.45	-	0	1.46

Median msg/s is the median sampled client-side throughput interval; it shows steady-state throughput without letting a short late-run stall dominate the whole-run average.

Rows and Comparison Ratio use Median msg/s when available; older result files without interval samples fall back to Messages/sec.

Drift compares last-third with first-third average throughput. Slope is the normalized least-squares trend; steady-state below 85% of peak or slope below -1%/min fails the regression gate.

Consumer (Batch) (15 minutes, 1000B messages, 16,384B seed batches)

Client	CPU μs/msg	CPU μs/request	Messages/sec	Median msg/s	Drift	Slope %/min	MB/sec	Accepted msg/s	Errors	Standing cores
Dekaf	1.15	-	1,198,245	1,195,583	-0.9%	-0.14%	1142.74	-	0	1.38

Median msg/s is the median sampled client-side throughput interval; it shows steady-state throughput without letting a short late-run stall dominate the whole-run average.

Rows and Comparison Ratio use Median msg/s when available; older result files without interval samples fall back to Messages/sec.

Drift compares last-third with first-third average throughput. Slope is the normalized least-squares trend; steady-state below 85% of peak or slope below -1%/min fails the regression gate.

Consumer (Raw Bytes) (15 minutes, 1000B messages, 16,384B seed batches)

Client	CPU μs/msg	CPU μs/request	Messages/sec	Median msg/s	Drift	Slope %/min	MB/sec	Accepted msg/s	Errors	Standing cores
Dekaf	0.43	-	3,624,160	3,647,900	+1.9%	+0.04%	3456.27	-	0	1.56

Median msg/s is the median sampled client-side throughput interval; it shows steady-state throughput without letting a short late-run stall dominate the whole-run average.

Rows and Comparison Ratio use Median msg/s when available; older result files without interval samples fall back to Messages/sec.

Drift compares last-third with first-third average throughput. Slope is the normalized least-squares trend; steady-state below 85% of peak or slope below -1%/min fails the regression gate.

Consumer (Raw Batch) (15 minutes, 1000B messages, 16,384B seed batches)

Client	CPU μs/msg	CPU μs/request	Messages/sec	Median msg/s	Drift	Slope %/min	MB/sec	Accepted msg/s	Errors	Standing cores
Dekaf	0.36	-	4,227,312	4,178,857	-0.6%	-0.11%	4031.48	-	0	1.53

Median msg/s is the median sampled client-side throughput interval; it shows steady-state throughput without letting a short late-run stall dominate the whole-run average.

Rows and Comparison Ratio use Median msg/s when available; older result files without interval samples fall back to Messages/sec.

Drift compares last-third with first-third average throughput. Slope is the normalized least-squares trend; steady-state below 85% of peak or slope below -1%/min fails the regression gate.

Memory & GC statistics — latest run

Client	Scenario	Gen0	Gen1	Gen2	Total Allocated	Alloc/msg
Confluent	Consumer	24027	103	0	2780.29 GB	2.38 KB
Confluent	Producer (Fire-and-Forget)	249305	1	1	1429.12 GB	1.26 KB
Confluent	Producer (Fire-and-Forget)	312378	1	1	1495.72 GB	1.26 KB
Confluent	Producer (Fire-and-Forget), 3 Brokers	206005	1	1	967.84 GB	1.26 KB
Confluent	Producer (Acks All)	309292	24	1	1485.71 GB	1.26 KB
Confluent	Producer (Acks All)	307843	1	1	1526.46 GB	1.26 KB
Confluent	Producer (Acks All), 3 Brokers	172292	15	1	834.72 GB	1.26 KB
Confluent	Producer (Fire-and-Forget, Idempotent)	310289	1	1	1524.28 GB	1.26 KB
Confluent	Producer (Fire-and-Forget, Idempotent)	323966	35	1	1560.23 GB	1.26 KB
Confluent	Producer (Fire-and-Forget, Idempotent), 3 Brokers	197379	1	1	927.40 GB	1.26 KB
Confluent	Producer → Consumer Round-Trip Steady State	5445	1	1	12.07 GB	655 B
Confluent	Producer (Transactional EOS), 3 Brokers	104	1	1	129.35 MB	863 B
Dekaf	Consumer	26426	17	2	3002.24 GB	1.98 KB
Dekaf	Consumer (Batch)	53452	3	1	2033.45 GB	1.98 KB
Dekaf	Consumer (Raw Bytes)	3	1	0	462.58 MB	0 B
Dekaf	Consumer (Raw Batch)	9	1	0	1.05 GB	0 B
Dekaf	Producer (Fire-and-Forget)	425	2	2	145.14 MB	0 B
Dekaf	Producer (Fire-and-Forget)	450	1	1	1.84 GB	1 B
Dekaf	Producer (Fire-and-Forget), 3 Brokers	218	3	2	913.14 MB	1 B
Dekaf	Producer (Acks All)	437	2	2	1.81 GB	1 B
Dekaf	Producer (Acks All)	425	2	2	98.31 MB	0 B
Dekaf	Producer (Acks All), 3 Brokers	159	3	2	702.40 MB	1 B
Dekaf	Producer (Fire-and-Forget, Idempotent)	401	2	2	121.07 MB	0 B
Dekaf	Producer (Fire-and-Forget, Idempotent)	423	2	2	1.61 GB	1 B
Dekaf	Producer (Fire-and-Forget, Idempotent), 3 Brokers	217	4	2	859.59 MB	1 B
Dekaf	Producer → Consumer Round-Trip Steady State	601	8	1	2.82 GB	153 B
Dekaf	Producer (Transactional EOS), 3 Brokers	194	8	1	486.53 MB	459 B
Dekaf (3conn)	Producer (Fire-and-Forget)	332	2	2	1.46 GB	1 B
Dekaf (3conn)	Producer (Fire-and-Forget), 3 Brokers	229	3	2	987.35 MB	1 B
Dekaf (3conn)	Producer (Fire-and-Forget, Idempotent)	303	2	2	1.25 GB	1 B
Dekaf (3conn)	Producer (Fire-and-Forget, Idempotent), 3 Brokers	241	3	2	995.14 MB	1 B

Confluent.Kafka uses native librdkafka; .NET GC allocation counters exclude unmanaged allocations.

---

About These Tests

Stress tests measure sustained performance over extended periods against real Kafka brokers, with both clients paired on the same VM for a fair comparison.

Methodology — how these numbers are produced

• Real Kafka: Tests run against actual Apache Kafka instances
• CPU Isolation: Brokers are pinned to dedicated cores and the client under test to its own cores, so the client — not the broker — is the measured bottleneck
• RAM-backed Broker Logs: Kafka log dirs are mounted on tmpfs so disk I/O never caps broker ingestion
• Delivered Throughput: producer tables report broker-confirmed throughput, measured as the end-offset delta across all partitions — not the client-side append rate, which can run far ahead of what the broker ever accepts
• Median Interval Throughput: table order and comparison ratios use median sampled client-side msg/s when available, which is less sensitive to short late-run stalls than the whole-run mean
• Same-VM Pairing: comparable Dekaf and Confluent scenarios run sequentially inside one job/VM; 1-broker producer acceptance lanes run twice in opposite client orders and publish a geometric-mean aggregate, while other lanes alternate order by workflow run number
• Backpressure Parity: both producers are bounded to the same 512 MB local buffer (Dekaf BufferMemory, librdkafka queue.buffering.max) and block on a full buffer, so neither client can absorb an unbounded backlog into RAM
• Consumer Loop Replay: Consumer tests re-read a pre-seeded topic (seek to beginning when drained) instead of racing a live feeder, so the consumer itself is measured; table headings report the 16KB seed batch size because it amplifies per-batch costs relative to well-batched workloads
• Delivery Latency Sampling: 1 in 1000 produced messages is awaited end-to-end to record true broker round-trip latency
• Round-Trip Correctness: Bounded sequenced payloads are consumed back and checked for corruption, wrong partitions, gaps, duplicates, and reordering
• Round-Trip CPU Scope: CPU time covers both bulk production and consumer validation; it is not a producer-only metric
• Round-Trip Alloc Scope: the GC/alloc window likewise spans production plus consume-side validation; values are deliberately consumed as byte[] on both clients for parity, so each consumed payload is materialized as a fresh array (~152 B at 128 B messages) — the expected allocation floor for this lane, not a leak
• CPU Efficiency: CPU time per message differentiates client efficiency even at equal throughput
• Noise-Aware Trends: each scenario is compared with its last 10 matching runs using a median ± 2×MAD band; one adverse excursion warns and two consecutive regressions fail the workflow
• Parallel Execution: Each scenario runs in its own isolated environment
• Both Clients: Direct comparison between Dekaf and Confluent.Kafka
• Memory Monitoring: Tracks GC behavior and memory usage over time
• Error Rates: Ensures stability under load