phpPHP Community
Community homeNews Release notes PHP wiki Web directory Multi-levelFeatured packages Best practices Q&A Jobs 38

PHP Community · Hiring: remote PHP engineer (full-time, long-term)

593
phpr/php·posted by zhou_yi·just nowTooling

Hiring: remote PHP engineer (full-time, long-term)

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.

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".

Image placeholder · object storage in production
POST /api/uploads → CDN origin pull
70 comments

70 comments

M
Ttang_haoOP·2 days ago

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

395
Zzhou_yi·2 days agoedited

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

422
Llinlin·12 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.

403
Rran_bo·just now

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

506
Rran_bo·28 minutes ago

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

123
Bbob_chen·28 minutes ago

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

79
Zzhou_yiMod·28 minutes ago

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

53
Sswoole_lee·1 hour agoeditedLevel 6

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.

147
Rrase·2 days agoLevel 6

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.

32
Wwinter·just now

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

17
Zzhou_yi·2 hours agoedited

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

314
Sslow_query·2 days agoeditedLevel 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.

32
Kkite·2 days ago

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

159
Wwinter·2 days ago

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

398
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.

347
Kkernel_panicOP·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.

287
Bbob_chen·2 days ago

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

310
Hhuang_ke·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.

291
Ddev_zhou·2 days agoedited

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

283
Sswoole_leeOP·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.

149
Kkernel_panic·12 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.

172
Kkite·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.

27
Nnikic·2 days ago

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

62
Ttang_hao·2 days ago

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

244
Lli_ming·2 days agoedited

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

42
Cchen_dev·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.

34
Rrase·2 days agoLevel 6

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

62
Hhuang_keOP·2 days agoeditedLevel 6

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.

2
Sslow_query·2 days agoLevel 6

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.

60
Lli_ming·3 minutes ago

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

3
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.

170
Lli_ming·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.

157
Ddev_zhou·2 days ago

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

7
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.

141
Ddev_zhou·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.

139
Rrase·2 days ago

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

116
Ddev_zhou·2 days ago

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

87
Nnikic·yesterday

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

77
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.

28
Mmike_xu·2 days ago

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

498
Rrase·2 days ago

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

17
Sswoole_lee·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.

149
Sswoole_lee·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.

76
Oops_wang·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.

68
Sswoole_lee·1 hour 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.

122
Mmike_xu·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.

114
Kkernel_panic·2 days ago

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

66
Zzhou_yi·yesterday

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

378
Sslow_query·12 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.

61
Kkernel_panic·yesterday

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.

40
Wwinter·2 days ago

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

420
Sslow_query·5 hours ago

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

32
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.

17
Wwinter·3 minutes 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.

6
Zzhou_yi·2 days agoLevel 6

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
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.

138
Aalice_devMod·2 days ago

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

84
Zzhu_zongMod·2 days ago

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

34
Oops_wang·2 days agoedited

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

32
Rran_boOP·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.

3
Nnikic·1 hour 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.

21
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.

154
Kkernel_panic·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.

382
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.

3
Kkite·12 minutes 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.

2
Oops_wang·5 hours ago

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

2
Rrase·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.

226
Zzhu_zong·3 minutes ago

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

2
Aalice_dev·2 days ago

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

1
WwinterMod·yesterday

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

1

This is the post detail page /en/c/php/post/p6. 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 →