ECEdge Computing · Showcase
Community homeHot discussions Latest posts Best of Knowledge base Resources About

Edge Computing · Showcase · The edge cases the edge-computing-show docs never spell out

134
ECr/edge-computing-show·posted by linlin·2 hours agoOpen source

The edge cases the edge-computing-show docs never spell out

Short version: edge-computing-show needs almost no tuning at small and medium scale — the point where it starts to hurt is much further out than most people assume. Full measurements below.

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.

-- The query that broke: a full scan over 20M rows.
-- A composite index took P99 from 1.8s down to 42ms.
SELECT id, title, created_at
  FROM posts
 WHERE community_id = ?
   AND status = 1
 ORDER BY score DESC
 LIMIT 20;

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.

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".

71 comments

71 comments

M
Llinlin·12 minutes ago

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

434
Kkernel_panic·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.

288
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
Ddev_zhou·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.

155
Llinlin·2 days ago

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

391
Rrase·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.

94
Rran_bo·2 days agoedited

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

386
Zzhou_yi·2 days agoedited

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.

332
Zzhou_yi·yesterdayedited

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

317
Zzhu_zong·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.

298
Kkite·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.

261
Lli_ming·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.

218
Sswoole_lee·3 minutes ago

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

184
Llinlin·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.

176
Wwinter·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.

122
Oops_wangMod·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.

16
Nnikic·2 days ago

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

102
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.

281
Aalice_devMod·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.

24
Rrase·3 minutes ago

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

200
Rrase·2 days ago

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

11
Nnikic·3 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.

181
Sslow_query·3 minutes ago

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

63
Wwinter·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.

379
Kkite·2 days agoedited

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

174
Oops_wang·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.

162
Aalice_devOP·3 minutes ago

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

94
Sswoole_lee·5 hours ago

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

136
Kkite·28 minutes agoedited

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.

129
Kkernel_panic·2 days ago

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

119
Hhuang_ke·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.

116
Kkernel_panic·2 hours ago

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

111
Sswoole_leeOP·3 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.

124
Mmike_xu·yesterday

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

60
Ttang_hao·2 days ago

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

288
Aalice_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.

94
Zzhou_yi·yesterday

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.

49
Rrase·5 hours agoedited

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

45
Ttang_hao·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.

11
Kkite·yesterday

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.

247
Bbob_chen·2 hours ago

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

156
Llinlin·2 days agoeditedLevel 6

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

444
Oops_wang·yesterday

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

3
Zzhou_yi·2 days ago

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

301
Wwinter·just now

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.

179
Nnikic·just nowedited

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

312
Oops_wang·2 days ago

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

22
Rran_bo·2 days ago

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

3
Kkernel_panic·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.

2
Rrase·2 days ago

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

513
KkiteOP·2 days ago

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

32
Aalice_devOP·yesterdayedited

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

6
Cchen_dev·yesterday

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

56
Zzhu_zong·2 days agoedited

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.

55
Kkite·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.

151
Mmike_xu·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.

304
Sslow_queryMod·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.

85
Wwinter·2 hours ago

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

56
WwinterOP·2 days ago

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

86
Zzhou_yi·2 days agoedited

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.

46
Ttang_hao·5 hours ago

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

35
Lli_ming·2 days ago

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

14
Aalice_dev·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.

29
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.

385
Nnikic·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.

369
Nnikic·2 days ago

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

24
Sswoole_lee·2 days ago

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

2
Sslow_query·3 minutes 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.

1
Sswoole_lee·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.

1
Zzhu_zong·yesterday

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.

69
Wwinter·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.

1

This is the post detail page /en/c/edge-computing-show/post/p0. 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 →