KAKafka · English
Community homeHot discussions Latest posts Best of Knowledge base Resources About

Kafka · English · The kafka-en execution flow in one diagram (with sequence chart)

310
KAr/kafka-en·posted by swoole_lee·just nowInternals

The kafka-en execution flow in one diagram (with sequence chart)

It took me two weeks of on-and-off digging and plenty of wrong turns. Writing the process down as it happened so the next person spends less time.

Order of investigation, by return on effort: 1. Check downstream latency first — usually it is not your problem 2. Then pool hit rate and wait-queue length 3. Only then GC and allocation 4. Suspect the framework last

What genuinely surprised me was the tail. The average looked great while P99 jumped by an order of magnitude past some threshold. The cause was not kafka-en itself but our upstream connection reuse — the load test traffic was too clean and hid the long-tail requests.

Worth noting: the official docs do cover this, just in a very inconspicuous spot. I only found it reading the source comments, where the author explains the reasoning — roughly "so that it degrades into predictable behaviour in extreme cases".

We also fixed monitoring along the way: replaced average-based alerts with percentiles and split them per endpoint. False alerts dropped by about seventy percent and the on-call rotation visibly cheered up.

One last trap: in container environments remember to adjust the memory-related parameters in step. Otherwise the host limit and the process expectation disagree, and the symptom is intermittent, unreproducible failure.

63 comments

63 comments

M
Hhuang_ke·1 hour ago

Can you give a minimal reproduction? I ran it locally for ten minutes and could not reproduce on macOS with the latest version.

425
Mmike_xu·2 days ago

We have run this in production for two years without hitting it. That said, we never reached this scale, so our experience is not really evidence here.

187
Rrase·2 days ago

There is actually a simpler fix that needs no architecture change: move this check up to the gateway and the problem disappears. The cost is one extra lookup at the gateway.

2
Aalice_dev·2 days ago

Sharing our numbers, 8 cores 16GB, same scenario:

| Concurrency | P50 | P99 |
|---|---|---|
| 200 | 12ms | 88ms |
| 500 | 31ms | 340ms |

P99 clearly collapses at 500 concurrency, which lines up with your knee point.

423
Sswoole_lee·2 days ago

One counter-example: below kafka-en 7.4 the semantics of that code are different, so do not copy it verbatim. We got burned in staging and rolled back once.

374
Mmike_xu·2 days ago

Thanks for sharing real numbers — far more useful than the articles that only cover concepts.

362
Nnikic·2 days ago

Can you give a minimal reproduction? I ran it locally for ten minutes and could not reproduce on macOS with the latest version.

338
RraseOP·2 days ago

Worth learning from this debugging approach. We went straight at the logs and took a much longer route.

207
Kkernel_panic·2 days ago

Agreeing with the above. One addition: with this option enabled the GC count in your metrics doubles, so adjust the alert threshold at the same time or it will keep firing.

252
Ddev_zhou·28 minutes ago

I just read the kafka-en source — the author actually explains the reasoning in a comment, roughly "so that it degrades into predictable behaviour in extreme cases".

245
Cchen_dev·3 minutes ago

Worth learning from this debugging approach. We went straight at the logs and took a much longer route.

195
LlinlinMod·3 minutes ago

This is not a kafka-en problem, it is a usage problem. The docs say this API is not thread-safe and you must lock around it yourself.

68
Lli_ming·28 minutes ago

I see point 3 differently. The trade-off depends on your read/write ratio: read-heavy with little writing means caching actually widens the inconsistency window.

449
Rran_bo·2 days ago

Agreeing with the above. One addition: with this option enabled the GC count in your metrics doubles, so adjust the alert threshold at the same time or it will keep firing.

494
Nnikic·2 days ago

A question: what changes in a container with a 512Mi memory limit? That is how we run it in production.

76
Oops_wang·2 days agoLevel 6

There is actually a simpler fix that needs no architecture change: move this check up to the gateway and the problem disappears. The cost is one extra lookup at the gateway.

6
Zzhu_zong·2 hours ago

Has anyone run a controlled experiment? I did, reducing it to a single variable, and the difference was 4% — within noise. So I suspect the main cause is something else.

257
Kkernel_panic·2 days ago

I see point 3 differently. The trade-off depends on your read/write ratio: read-heavy with little writing means caching actually widens the inconsistency window.

317
Ddev_zhou·12 minutes agoedited

Thanks for sharing real numbers — far more useful than the articles that only cover concepts.

126
LlinlinOP·2 days agoLevel 6

I just read the kafka-en source — the author actually explains the reasoning in a comment, roughly "so that it degrades into predictable behaviour in extreme cases".

2
Nnikic·2 days ago

I see point 3 differently. The trade-off depends on your read/write ratio: read-heavy with little writing means caching actually widens the inconsistency window.

42
Aalice_dev·1 hour ago

This matches what we see in production. We only hit it past 3k QPS; the earlier load tests showed nothing — the test traffic was too clean, with no long-tail requests.

22
Kkernel_panic·2 days agoLevel 6

There is actually a simpler fix that needs no architecture change: move this check up to the gateway and the problem disappears. The cost is one extra lookup at the gateway.

247
Aalice_dev·12 minutes agoLevel 6

A question: what changes in a container with a 512Mi memory limit? That is how we run it in production.

7
Oops_wang·2 hours ago

We have run this in production for two years without hitting it. That said, we never reached this scale, so our experience is not really evidence here.

8
Wwinter·yesterdayLevel 6

There is actually a simpler fix that needs no architecture change: move this check up to the gateway and the problem disappears. The cost is one extra lookup at the gateway.

45
Lli_ming·2 days agoLevel 6

Worth learning from this debugging approach. We went straight at the logs and took a much longer route.

35
Nnikic·yesterday

One counter-example: below kafka-en 7.4 the semantics of that code are different, so do not copy it verbatim. We got burned in staging and rolled back once.

253
Llinlin·3 minutes ago

I just read the kafka-en source — the author actually explains the reasoning in a comment, roughly "so that it degrades into predictable behaviour in extreme cases".

242
Rran_bo·2 hours ago

Saved. I am reworking this area this week — this saves a lot of wrong turns.

4
Rran_bo·2 hours ago

This matches what we see in production. We only hit it past 3k QPS; the earlier load tests showed nothing — the test traffic was too clean, with no long-tail requests.

501
Ddev_zhou·2 days agoLevel 6

Saved. I am reworking this area this week — this saves a lot of wrong turns.

11
Ddev_zhouOPMod·2 days ago

Can you give a minimal reproduction? I ran it locally for ten minutes and could not reproduce on macOS with the latest version.

347
Wwinter·2 days ago

Has anyone run a controlled experiment? I did, reducing it to a single variable, and the difference was 4% — within noise. So I suspect the main cause is something else.

88
Rrase·1 hour ago

Thanks for sharing real numbers — far more useful than the articles that only cover concepts.

144
Cchen_devOP·2 days ago

One counter-example: below kafka-en 7.4 the semantics of that code are different, so do not copy it verbatim. We got burned in staging and rolled back once.

163
Hhuang_ke·just now

I see point 3 differently. The trade-off depends on your read/write ratio: read-heavy with little writing means caching actually widens the inconsistency window.

96
Llinlin·2 days ago

This is not a kafka-en problem, it is a usage problem. The docs say this API is not thread-safe and you must lock around it yourself.

193
Bbob_chen·2 days agoedited

Thanks for sharing real numbers — far more useful than the articles that only cover concepts.

30
Aalice_dev·2 days ago

Can you give a minimal reproduction? I ran it locally for ten minutes and could not reproduce on macOS with the latest version.

24
Zzhou_yi·2 days ago

We have run this in production for two years without hitting it. That said, we never reached this scale, so our experience is not really evidence here.

34
Sswoole_lee·yesterday

Saved. I am reworking this area this week — this saves a lot of wrong turns.

185
Hhuang_ke·2 days ago

A question: what changes in a container with a 512Mi memory limit? That is how we run it in production.

228
Kkernel_panic·2 days ago

I just read the kafka-en source — the author actually explains the reasoning in a comment, roughly "so that it degrades into predictable behaviour in extreme cases".

25
Aalice_dev·5 hours agoedited

This is not a kafka-en problem, it is a usage problem. The docs say this API is not thread-safe and you must lock around it yourself.

167
Lli_mingOP·2 days ago

This matches what we see in production. We only hit it past 3k QPS; the earlier load tests showed nothing — the test traffic was too clean, with no long-tail requests.

378
Cchen_dev·2 days agoedited

This matches what we see in production. We only hit it past 3k QPS; the earlier load tests showed nothing — the test traffic was too clean, with no long-tail requests.

92
Mmike_xu·2 days ago

Saved. I am reworking this area this week — this saves a lot of wrong turns.

88
Lli_ming·2 days ago

One counter-example: below kafka-en 7.4 the semantics of that code are different, so do not copy it verbatim. We got burned in staging and rolled back once.

163
Rrase·2 days ago

Saved. I am reworking this area this week — this saves a lot of wrong turns.

123
Rran_bo·2 days ago

This is not a kafka-en problem, it is a usage problem. The docs say this API is not thread-safe and you must lock around it yourself.

107
Zzhou_yi·2 hours ago

Agreeing with the above. One addition: with this option enabled the GC count in your metrics doubles, so adjust the alert threshold at the same time or it will keep firing.

95
Sslow_query·2 hours ago

Worth learning from this debugging approach. We went straight at the logs and took a much longer route.

90
Sswoole_lee·2 hours ago

Has anyone run a controlled experiment? I did, reducing it to a single variable, and the difference was 4% — within noise. So I suspect the main cause is something else.

64
Bbob_chen·2 days ago

Sharing our numbers, 8 cores 16GB, same scenario:

| Concurrency | P50 | P99 |
|---|---|---|
| 200 | 12ms | 88ms |
| 500 | 31ms | 340ms |

P99 clearly collapses at 500 concurrency, which lines up with your knee point.

1
Sswoole_lee·2 days ago

Agreeing with the above. One addition: with this option enabled the GC count in your metrics doubles, so adjust the alert threshold at the same time or it will keep firing.

4
Aalice_dev·2 days ago

Has anyone run a controlled experiment? I did, reducing it to a single variable, and the difference was 4% — within noise. So I suspect the main cause is something else.

51
Oops_wang·2 days ago

Sharing our numbers, 8 cores 16GB, same scenario:

| Concurrency | P50 | P99 |
|---|---|---|
| 200 | 12ms | 88ms |
| 500 | 31ms | 340ms |

P99 clearly collapses at 500 concurrency, which lines up with your knee point.

35
Zzhu_zong·2 days ago

A question: what changes in a container with a 512Mi memory limit? That is how we run it in production.

31
Llinlin·12 minutes ago

Sharing our numbers, 8 cores 16GB, same scenario:

| Concurrency | P50 | P99 |
|---|---|---|
| 200 | 12ms | 88ms |
| 500 | 31ms | 340ms |

P99 clearly collapses at 500 concurrency, which lines up with your knee point.

29
Llinlin·2 days ago

Agreeing with the above. One addition: with this option enabled the GC count in your metrics doubles, so adjust the alert threshold at the same time or it will keep firing.

28
Bbob_chen·2 days ago

We have run this in production for two years without hitting it. That said, we never reached this scale, so our experience is not really evidence here.

2
Zzhu_zong·2 days ago

One counter-example: below kafka-en 7.4 the semantics of that code are different, so do not copy it verbatim. We got burned in staging and rolled back once.

132

This is the post detail page /en/c/kafka-en/post/p10. Posts and comments are generated deterministically from a seeded PRNG, so the same post always renders the same content and the link can be shared, reloaded and indexed. In production this page reads MySQL for the post, Redis for hot-post caching, and fetches the whole comment tree in a single query on the path column.

See the database schema →