hetzner / fsn-1 — CPX32

€42.23/mo · storage net-fast · host afad71924d76 · 2026-07-17T15:22:40Z · submitter uptimeify

Profiles

profile	grade	bound by
postgres_oltp	C	`disk.wal_fsync.p999_us`
timescale_ingest	C	`disk.wal_fsync.p999_us`
patroni_member	F	`cpu.single_thread_eps`
redis_sentinel	F	`cpu.single_thread_eps`
worker_probe	F	`cpu.single_thread_eps`
playwright_node	F	`cpu.single_thread_eps`
nuxt_ssr	F	`cpu.single_thread_eps`

Tail-latency reality

fsync cumulative distribution

hover the curve — snaps to measured percentiles

Points are the measured p50 / p99 / p99.9 / max — the line between them is interpolated (we store percentiles, not a full histogram). Fixed axis: a tight curve on the left is a good host; a curve swept right is an exploding tail.

marketing claim vs measured

− typical marketing claim + what p99bench measured here

−"Up to N,000 IOPS"

+ fsync p99.9 latency at QD1 4.42 ms C

−"NVMe SSD — low latency"

+ p99.9 vs median, not the mean p50 0.24 ms → p99.9 1.06 ms A

−"Dedicated vCPUs"

+ CPU steal measured under full load 0% steal

−"High-performance cores"

+ single-thread throughput 333 eps F

−"Fast" (short benchmark)

+ IOPS drop over 30 min sustained load 1.84% degradation

−"No noisy neighbours"

+ scheduler stall p99.9 0.7 ms C

Disk

bound by disk.wal_fsync.p999_us

WAL fsync latency ladder (ms, QD1) Every COMMIT waits on one fdatasync. p99.9 is the graded metric — the tail that stalls a database.

avg

0.93 ms

p50

0.86 ms

p99

1.8 ms

p99.9

4.4 ms C

max

12.7 ms

8K random latency p99.9 by operation (ms)

read QD1

1.1 ms

read

8.2 ms

write

30.3 ms

mixed

8.8 ms

Throughput

seq read

5,267 MB/s

seq write

1,950 MB/s

Random IOPS

read 8K

58,657 IOPS

write 8K

40,883 IOPS

mixed 8K

36,529 IOPS

fsync

965 IOPS

Sustained IOPS — first vs last minute of 30 min 1.84% degradation over 30 min

first min

51,196 IOPS

last min

50,255 IOPS

exact values & per-metric grades

metric	value	grade
`disk.wal_fsync.p999_us`	4423.68	C
`disk.wal_fsync.iops`	964.78	C
`disk.rand_read_8k_qd1.p99_us`	464.9	A
`disk.rand_read_8k.iops`	58657.04	B
`disk.rand_write_8k.iops`	40882.71	B
`disk.seq_write.bw_mbs`	1949.59	A
`disk.seq_read.bw_mbs`	5267.15	A
`disk.steady_state.degradation_pct`	1.84	A

CPU

bound by cpu.single_thread_eps

Throughput (events/s)

single

333 eps F

multi

1,227 eps

Scheduler stall ladder (ms) How long the kernel keeps a runnable thread waiting. p99.9 is graded.

p99

0.19 ms

p99.9

0.7 ms C

max

11.1 ms

Sustained throughput — first vs last minute 2% degradation · steal 0%

first min

1,323 eps

last min

1,290 eps

exact values & per-metric grades

metric	value	grade
`cpu.single_thread_eps`	332.9	F
`cpu.scaling_efficiency`	0.921	A
`cpu.steal_pct_under_load`	0	A
`cpu.stall_p999_us`	698	C
`cpu.steady_state.degradation_pct`	2	A
`cpu.tls_verify_s`	11201.3	C

RAM

bound by ram.bw_read_mbs

Memory bandwidth (MB/s) bw read is measured outside cache (the graded metric); seq/rnd are raw STREAM-style figures.

bw read

37,556 MB/s B

seq read

73,262 MB/s

seq write

44,010 MB/s

rnd read

7,515 MB/s

rnd write

685 MB/s

exact values & per-metric grades

metric	value	grade
`ram.bw_read_mbs`	37555.89	B

Network

bound by network.rtt_jitter_ratio

Round-trip p99 by target (ms) Same fixed target list from every host — distance is constant, so peering and neighbour noise surface.

hetzner-fsn1

2.2 ms

hetzner-hel1

25.8 ms

hetzner-ash

95.3 ms

Packet loss by target (%)

hetzner-fsn1

0 %

hetzner-hel1

0 %

hetzner-ash

0 %

exact values & per-metric grades

metric	value	grade
`network.loss_pct`	0	A
`network.rtt_jitter_ratio`	4.44	D