upcloud / de-fra1 — CLOUDNATIVE-2xCPU-4GB-storage-standard-25GB

€20.59/mo · storage net-fast · host 7daa167c9114 · 2026-07-19T10:42:32Z · 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	D	`network.rtt_jitter_ratio`
playwright_node	D	`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 21.6 ms D

−"NVMe SSD — low latency"

+ p99.9 vs median, not the mean p50 0.07 ms → p99.9 9.63 ms A

−"Dedicated vCPUs"

+ CPU steal measured under full load 0.1% steal

−"High-performance cores"

+ single-thread throughput 2,194 eps A

−"Fast" (short benchmark)

+ IOPS drop over 30 min sustained load 0% degradation

−"No noisy neighbours"

+ scheduler stall p99.9 0.11 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

0.62 ms

p50

0.47 ms

p99

1.8 ms

p99.9

21.6 ms D

max

57.4 ms

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

read QD1

9.6 ms

read

103 ms

write

1,955 ms

mixed

1,116 ms

Throughput

seq read

400 MB/s

seq write

400 MB/s

Random IOPS

read 8K

10,002 IOPS

write 8K

7,354 IOPS

mixed 8K

8,143 IOPS

fsync

1,541 IOPS

Sustained IOPS — first vs last minute of 30 min 0% degradation over 30 min

first min

12,706 IOPS

last min

13,391 IOPS

exact values & per-metric grades

metric	value	grade
`disk.wal_fsync.p999_us`	21626.88	D
`disk.wal_fsync.iops`	1540.51	B
`disk.rand_read_8k_qd1.p99_us`	248.83	A
`disk.rand_read_8k.iops`	10001.83	D
`disk.rand_write_8k.iops`	7354.4	D
`disk.seq_write.bw_mbs`	399.74	C
`disk.seq_read.bw_mbs`	400.39	D
`disk.steady_state.degradation_pct`	0	A

CPU

bound by cpu.stall_p999_us

Throughput (events/s)

single

2,194 eps A

multi

4,296 eps

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

p99

0.07 ms

p99.9

0.11 ms B

max

3.7 ms

Sustained throughput — first vs last minute 0% degradation · steal 0.05%

first min

4,371 eps

last min

4,393 eps

exact values & per-metric grades

metric	value	grade
`cpu.single_thread_eps`	2194.21	A
`cpu.scaling_efficiency`	0.978	A
`cpu.steal_pct_under_load`	0.1	A
`cpu.stall_p999_us`	109	B
`cpu.steady_state.degradation_pct`	0	A
`cpu.tls_verify_s`	26276.3	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

72,549 MB/s A

seq read

169,174 MB/s

seq write

63,324 MB/s

rnd read

12,514 MB/s

rnd write

3,938 MB/s

exact values & per-metric grades

metric	value	grade
`ram.bw_read_mbs`	72548.69	A

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

11.8 ms

hetzner-hel1

23 ms

hetzner-ash

91.4 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`	2.3274161735700196	D