Polars · Beginners · Hiring: remote polars-newbie engineer (full-time, long-term)

274
POr/polars-newbie·posted by bob_chen·yesterdayAnnouncement

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

Short version: polars-newbie 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.

// Minimal reproduction: you must use a real long-tail distribution here.
// Uniform load-test traffic will never trigger this.
func (s *Server) handle(ctx context.Context) error {
    conn, err := s.pool.Acquire(ctx)
    if err != nil {
        return fmt.Errorf("acquire: %w", err)
    }
    defer conn.Release()

    return s.do(ctx, conn)
}

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.

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 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.

48 comments

48 comments

M
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.

221
Wwinter·28 minutes ago

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

204
Nnikic·5 hours ago

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

196
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.

196
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.

193
Kkite·2 days ago

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

118
Mmike_xu·2 days ago

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

162
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.

386
Sslow_query·12 minutes ago

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

160
Ddev_zhou·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.

123
Sslow_query·2 days ago

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

45
Zzhou_yi·yesterday

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

15
Aalice_dev·just now

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

12
Ttang_hao·12 minutes ago

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

58
Ttang_hao·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.

7
Sslow_query·just now

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

6
Rrase·12 minutes ago

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

1
Cchen_dev·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.

431
Zzhou_yi·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.

452
Sswoole_lee·5 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.

423
Llinlin·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.

136
Hhuang_ke·2 days ago

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

84
Sslow_query·5 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.

1
Kkernel_panic·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.

288
Rran_bo·2 days agoeditedLevel 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.

219
Rran_bo·2 hours ago

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

31
Ddev_zhou·12 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.

385
Kkernel_panic·1 hour ago

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

302
Ttang_haoMod·28 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.

117
Zzhu_zong·28 minutes agoedited

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.

4
Zzhou_yi·2 days ago

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

55
Kkernel_panic·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.

149
Cchen_devOP·2 days ago

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

52
Ddev_zhou·28 minutes ago

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

14
Bbob_chen·just now

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

442
Kkite·28 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.

304
Zzhou_yi·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
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.

1
Nnikic·3 minutes ago

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

1
Bbob_chenOP·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.

1
Kkernel_panic·2 days ago

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

114
Zzhou_yi·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.

13
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.

42
Oops_wang·12 minutes ago

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

13
Kkite·12 minutes ago

One counter-example: below polars-newbie 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
Oops_wang·2 days agoedited

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

16
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.

377
Rrase·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.

107

This is the post detail page /en/c/polars-newbie/post/p13. 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 →