ovh / prg — vps-1-lz-2026

€7.49/mo · storage net-slow · host 1d0480bc0fe6 · 2026-07-17T15:22:29Z · submitter uptimeify

Profiles

profile	grade	bound by
postgres_oltp	D	`disk.wal_fsync.p999_us`
timescale_ingest	D	`disk.wal_fsync.p999_us`
patroni_member	D	`disk.wal_fsync.p999_us`
redis_sentinel	D	`disk.wal_fsync.p999_us`
worker_probe	B	`cpu.stall_p999_us`
playwright_node	C	`host.ram_mb`
nuxt_ssr	B	`cpu.stall_p999_us`

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 15.4 ms D

−"NVMe SSD — low latency"

+ p99.9 vs median, not the mean p50 0.06 ms → p99.9 1.2 ms B

−"Dedicated vCPUs"

+ CPU steal measured under full load 0.02% steal

−"High-performance cores"

+ single-thread throughput 1,620 eps A

−"Fast" (short benchmark)

+ IOPS drop over 30 min sustained load 0.36% degradation

−"No noisy neighbours"

+ scheduler stall p99.9 0.35 ms B

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

3.5 ms

p50

3.1 ms

p99

9.5 ms

p99.9

15.4 ms D

max

63.5 ms

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

read QD1

1.2 ms

read

25.3 ms

write

1,468 ms

mixed

24 ms

Throughput

seq read

301 MB/s

seq write

300 MB/s

Random IOPS

read 8K

7,512 IOPS

write 8K

6,633 IOPS

mixed 8K

7,512 IOPS

fsync

278 IOPS

Sustained IOPS — first vs last minute of 30 min 0.36% degradation over 30 min

first min

10,744 IOPS

last min

10,706 IOPS

exact values & per-metric grades

metric	value	grade
`disk.wal_fsync.p999_us`	15400.96	D
`disk.wal_fsync.iops`	277.86	D
`disk.rand_read_8k_qd1.p99_us`	1122.3	B
`disk.rand_read_8k.iops`	7512.42	D
`disk.rand_write_8k.iops`	6633.33	D
`disk.seq_write.bw_mbs`	300.46	C
`disk.seq_read.bw_mbs`	300.51	D
`disk.steady_state.degradation_pct`	0.36	A

CPU

bound by cpu.stall_p999_us

Throughput (events/s)

single

1,620 eps A

multi

6,243 eps

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

p99

0.12 ms

p99.9

0.35 ms B

max

3.4 ms

Sustained throughput — first vs last minute 0% degradation · steal 0.03%

first min

6,277 eps

last min

6,312 eps

exact values & per-metric grades

metric	value	grade
`cpu.single_thread_eps`	1620.08	A
`cpu.scaling_efficiency`	0.963	A
`cpu.steal_pct_under_load`	0.02	A
`cpu.stall_p999_us`	354	B
`cpu.steady_state.degradation_pct`	0	A
`cpu.tls_verify_s`	15566	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

65,813 MB/s A

seq read

201,345 MB/s

seq write

56,469 MB/s

rnd read

13,451 MB/s

rnd write

3,522 MB/s

exact values & per-metric grades

metric	value	grade
`ram.bw_read_mbs`	65812.85	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

14.6 ms

hetzner-hel1

29.7 ms

hetzner-ash

98 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`	1.0279329608938548	A