ovh / eu-south-mil-b — c3-8

€65.52/mo · storage net-fast · host b73916d1f9cc · 2026-07-22T16:06:08Z · submitter uptimeify

Profiles

profile	grade	bound by
postgres_oltp	C	`disk.rand_read_8k.iops`
timescale_ingest	B	`disk.rand_write_8k.iops`
patroni_member	A	`disk.wal_fsync.p999_us`
redis_sentinel	A	`cpu.stall_p999_us`
worker_probe	B	`cpu.tls_verify_s`
playwright_node	C	`host.ram_mb`
nuxt_ssr	A	`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 0.67 ms A

−"NVMe SSD — low latency"

+ p99.9 vs median, not the mean p50 0.15 ms → p99.9 0.26 ms A

−"Dedicated vCPUs"

+ CPU steal measured under full load 0.01% steal

−"High-performance cores"

+ single-thread throughput 1,681 eps A

−"Fast" (short benchmark)

+ IOPS drop over 30 min sustained load 0.3% degradation

−"No noisy neighbours"

+ scheduler stall p99.9 0.07 ms A

Disk

bound by disk.rand_read_8k.iops

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

0.31 ms

p50

0.36 ms

p99

0.42 ms

p99.9

0.67 ms A

max

6.4 ms

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

read QD1

0.26 ms

read

3.5 ms

write

17.4 ms

mixed

3.6 ms

Throughput

seq read

7,150 MB/s

seq write

1,966 MB/s

Random IOPS

read 8K

40,065 IOPS

write 8K

23,464 IOPS

mixed 8K

40,066 IOPS

fsync

2,903 IOPS

Sustained IOPS — first vs last minute of 30 min 0.3% degradation over 30 min

first min

57,309 IOPS

last min

57,135 IOPS

exact values & per-metric grades

metric	value	grade
`disk.wal_fsync.p999_us`	667.65	A
`disk.wal_fsync.iops`	2902.88	B
`disk.rand_read_8k_qd1.p99_us`	158.72	A
`disk.rand_read_8k.iops`	40064.91	C
`disk.rand_write_8k.iops`	23464.04	B
`disk.seq_write.bw_mbs`	1965.6	A
`disk.seq_read.bw_mbs`	7150.07	A
`disk.steady_state.degradation_pct`	0.3	A

CPU

bound by cpu.tls_verify_s

Throughput (events/s)

single

1,681 eps A

multi

6,719 eps

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

p99

0.07 ms

p99.9

0.07 ms A

max

4 ms

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

first min

6,724 eps

last min

6,722 eps

exact values & per-metric grades

metric	value	grade
`cpu.single_thread_eps`	1680.68	A
`cpu.scaling_efficiency`	0.999	A
`cpu.steal_pct_under_load`	0.01	A
`cpu.stall_p999_us`	74	A
`cpu.steady_state.degradation_pct`	0	A
`cpu.tls_verify_s`	16414.3	B

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

77,422 MB/s A

seq read

219,918 MB/s

seq write

11,697 MB/s

rnd read

16,574 MB/s

rnd write

882 MB/s

exact values & per-metric grades

metric	value	grade
`ram.bw_read_mbs`	77421.94	A

Network

bound by network.loss_pct

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

hetzner-fsn1

14.3 ms

hetzner-hel1

29.4 ms

hetzner-ash

92.5 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`	1.0070422535211268	A