NINixOS · Toolchain
Community homeHot discussions Latest posts Best of Knowledge base Resources About

NixOS · Toolchain · War story: nixos-tools timezone handling almost cost us a full day of billing

98
NIr/nixos-tools·posted by tang_hao·3 days agoOpen sourceLocked

War story: nixos-tools timezone handling almost cost us a full day of billing

Most nixos-tools articles stop at "how to use it" and never cover "when not to use it". This is an attempt at the second half.

# 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

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 nixos-tools 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".

41 comments

41 comments

M
Rrase·1 hour 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.

497
Mmike_xu·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.

392
Lli_ming·1 hour ago

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

229
Lli_ming·yesterday

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.

182
Sslow_query·1 hour ago

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

208
Ttang_haoOP·2 days ago

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

105
KkiteOP·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
Nnikic·2 hours 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.

371
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.

207
Kkite·2 days ago

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

12
Oops_wangMod·2 hours ago

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

190
Rrase·2 days ago

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

179
Cchen_dev·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.

169
Nnikic·1 hour 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.

20
Kkite·2 days agoedited

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

42
Lli_ming·2 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.

167
Cchen_dev·3 minutes ago

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

101
Kkite·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.

160
Cchen_dev·just now

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

488
Rran_bo·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.

50
Cchen_dev·just nowedited

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.

132
Kkite·12 minutes ago

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

305
Ttang_hao·3 minutes agoedited

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

369
Lli_ming·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.

63
Oops_wang·2 days ago

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

2
Aalice_dev·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.

353
Ttang_hao·3 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.

361
Rrase·5 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.

348
Bbob_chen·2 days ago

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

213
Wwinter·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.

261
RraseOP·2 days ago

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

25
Cchen_dev·28 minutes agoedited

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

70
Oops_wang·28 minutes ago

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

404
Nnikic·just nowedited

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.

100
Lli_mingOP·yesterday

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

2
Ttang_hao·2 days ago

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

36
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.

18
KkiteMod·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.

14
Kkite·2 days ago

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

80
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.

4
Sswoole_lee·2 days agoedited

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

9

This is the post detail page /en/c/nixos-tools/post/p9. 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 →