VVue Community
Community homeEcosystem Release news Component libraries Q&A

Vue Community · This Vue bug hid for three years before anyone reported it

2.4K
Vr/vue·posted by huang_ke·yesterdayTranslation

This Vue bug hid for three years before anyone reported it

Some background first. Our setup is Vue plus three downstream services, seven figures of daily requests, peaking around nine in the evening.

The first thing was to collapse the variables. We were changing config and upgrading the version at the same time, and afterwards nobody could say which change caused what. We rolled back to moving one variable at a time, re-ran three times, and only then did the curve settle. Tedious, but not skippable.

# Load test: do not jump straight to the max concurrency.
# Ramp it up, otherwise you miss the knee.
for c in 50 100 200 400 800; do
  wrk -t8 -c$c -d60s --latency http://127.0.0.1:8080/api/feed
  sleep 20
done

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.

On trade-offs, my view is this: if nobody on the team owns this area long-term, do not introduce a second mechanism. With two coexistence you first have to work out which one is even in play when things break, and that costs far more than the performance you saved.

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 Vue itself but our upstream connection reuse — the load test traffic was too clean and hid the long-tail requests.

1086 comments

1086 comments

· first 120 loaded
M
Llinlin·5 hours 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.

520
Zzhu_zong·5 hours agoedited

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

23
Oops_wang·2 days ago

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

437
Zzhu_zong·2 days ago

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

1
Bbob_chen·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.

429
Lli_ming·2 hours ago

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

512
Wwinter·12 minutes 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.

299
Mmike_xu·2 days ago

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

162
Kkernel_panicOP·2 days ago

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

1
Nnikic·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.

415
Wwinter·12 minutes ago

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

20
Ddev_zhou·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.

364
Nnikic·2 days ago

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

46
Ddev_zhou·3 minutes agoedited

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.

2
Rrase·3 minutes ago

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

18
Kkernel_panic·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.

342
Lli_ming·2 days ago

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

422
Ttang_hao·12 minutes 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.

170
Ddev_zhou·2 days ago

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

465
Kkite·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.

11
Nnikic·2 days ago

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

322
Rran_boMod·28 minutes ago

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

321
Kkite·2 hours 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.

489
Wwinter·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.

305
Lli_ming·2 hours 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.

45
Ttang_haoOP·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.

88
Kkite·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.

24
Kkite·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.

26
Wwinter·2 days agoLevel 6

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

28
Nnikic·3 minutes 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.

1
Oops_wang·yesterday

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

1
Hhuang_ke·28 minutes 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.

14
Rrase·2 days agoedited

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

304
Zzhou_yiOP·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.

18
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.

249
Zzhu_zong·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.

217
Cchen_dev·1 hour ago

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

61
Nnikic·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.

23
Nnikic·2 days ago

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

213
Wwinter·2 days agoedited

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.

205
Rrase·2 days ago

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

198
LlinlinOP·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.

123
Cchen_dev·2 days ago

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

171
Kkernel_panic·just now

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

155
Zzhu_zong·5 hours ago

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

7
Oops_wang·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.

141
Llinlin·2 days ago

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

65
Lli_ming·2 days ago

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

1
Ttang_hao·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.

95
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.

92
Ddev_zhou·5 hours ago

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

1
Kkite·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.

90
Llinlin·yesterday

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

89
Lli_ming·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.

237
Ttang_hao·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.

86
Kkernel_panic·5 hours 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.

72
Cchen_dev·2 days ago

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

109
Ttang_hao·1 hour ago

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

65
Zzhou_yi·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.

59
Sswoole_lee·2 days ago

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

26
Ddev_zhou·1 hour 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.

48
Kkernel_panic·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.

47
Kkite·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.

45
Hhuang_ke·28 minutes 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.

33
Bbob_chen·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.

33
Ddev_zhou·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.

264
Cchen_dev·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.

33
Lli_ming·12 minutes ago

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

27
Aalice_dev·28 minutes ago

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

26
Llinlin·1 hour ago

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

355
Lli_ming·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.

8
Rrase·12 minutes ago

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

206
Bbob_chen·2 days ago

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

476
Hhuang_ke·2 days agoedited

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

2
WwinterMod·2 days ago

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

264
Ddev_zhouOP·2 days ago

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

33
Kkite·5 hours ago

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

40
Ddev_zhou·5 hours agoedited

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

95
Nnikic·28 minutes ago

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

451
Rran_bo·just now

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.

371
Zzhu_zongOP·2 hours ago

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

15
Rran_boMod·3 minutes ago

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

1
Wwinter·2 days ago

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

227
Kkernel_panic·just nowLevel 6

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

318
Sslow_query·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.

19
LlinlinMod·2 days agoLevel 6

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.

93
Hhuang_ke·2 days agoLevel 6

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

67
Nnikic·2 hours agoLevel 6

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

52
Mmike_xu·2 days agoLevel 6

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

15
Mmike_xu·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.

285
Zzhu_zong·2 days ago

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

403
Bbob_chenOP·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.

7
Kkite·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.

11
Ttang_haoMod·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.

220
Sslow_query·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.

68
Zzhu_zongMod·2 days agoedited

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

25
Llinlin·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.

23
Kkernel_panic·1 hour agoedited

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

103
Llinlin·12 minutes ago

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

21
Mmike_xu·2 days ago

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

18
Zzhou_yi·2 days ago

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

11
Cchen_dev·12 minutes ago

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

10
Zzhou_yi·2 days ago

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

51
Aalice_devOP·yesterday

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

221
Lli_ming·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.

40
Mmike_xu·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.

10
Hhuang_ke·2 days ago

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

9
Mmike_xu·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.

74
LlinlinOP·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.

481
Bbob_chen·3 minutes 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.

6
Cchen_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.

3
Aalice_dev·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.

2
NnikicMod·2 days ago

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

2
Sswoole_lee·2 days ago

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

2
Rrase·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.

273
Nnikic·2 days ago

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

2
Bbob_chen·2 days agoedited

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.

1
Kkite·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.

167
Ddev_zhou·2 days agoedited

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

1
Rrase·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.

45

This is the post detail page /en/c/vue/post/p12. 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 →