A four-state parallel coordinates plot for mixed-unit records — per-axis normalisation, drag-to-brush filtering that dims rather than deletes, draggable axis order, and a rank-correlation reading for every neighbouring pair.
Preview in your theme
Loading preview…
This component is part of Pro
The install command, prompt and concept map are included with a Pro license — one purchase unlocks every Pro component, forever.
status="ready" · 28 records × 6 axes. Each axis is normalised over its own [min, max], which is what lets milliseconds, dollars and instance counts share one frame; latency, error rate and cost are marked invert in the contract, so the good end is always up.
Service fleet · last 24 h
Six mixed-unit axes · latency, error rate and cost are drawn upside down so high is always good
EU-west
US-east
AP-south
28 of 28 shown
Point at a line to read it, or Tab into the plot and use the arrow keys. Drag on an axis to filter, drag a header to reorder.
Strongest neighbouring pair: Cost / 1k req and Throughput, rank correlation +0.99 over 28 shown records — their segments cross heavily.
Service fleet · last 24 h: every record with its value on every axis, in the current axis order. No filter is active, so every record counts as shown.
Record
Group
p95 latency (ms)
Error rate (%)
Cost / 1k req (USD)
Throughput (req/s)
CPU headroom (%)
Replicas
In filter
pixel-1
EU-west
90 ms
0.19%
0.351 USD
669 req/s
30%
19
yes
driver-2
US-east
126 ms
0.44%
0.337 USD
661 req/s
20%
16
yes
sensor-3
AP-south
108 ms
0.53%
0.315 USD
640 req/s
20%
15
yes
port-4
EU-west
113 ms
0.72%
0.114 USD
229 req/s
19%
5
yes
capacitor-5
US-east
105 ms
0.37%
0.283 USD
515 req/s
43%
16
yes
driver-6
AP-south
146 ms
0.8%
0.378 USD
727 req/s
35%
18
yes
program-7
EU-west
164 ms
0.76%
0.314 USD
559 req/s
45%
17
yes
sensor-8
US-east
131 ms
0.66%
0.223 USD
412 req/s
39%
11
yes
card-9
AP-south
164 ms
0.87%
0.214 USD
390 req/s
26%
10
yes
monitor-10
EU-west
149 ms
0.86%
0.137 USD
204 req/s
55%
8
yes
microchip-11
US-east
109 ms
0.66%
0.222 USD
458 req/s
23%
10
yes
port-12
AP-south
167 ms
0.86%
0.426 USD
786 req/s
44%
22
yes
interface-13
EU-west
96 ms
0.32%
0.162 USD
295 req/s
38%
9
yes
port-14
US-east
149 ms
0.85%
0.401 USD
774 req/s
29%
19
yes
monitor-15
AP-south
146 ms
0.86%
0.394 USD
735 req/s
39%
20
yes
bus-16
EU-west
103 ms
0.41%
0.046 USD
55 req/s
74%
3
yes
application-17
US-east
122 ms
0.42%
0.401 USD
768 req/s
39%
21
yes
firewall-18
AP-south
87 ms
0.11%
0.394 USD
764 req/s
31%
21
yes
card-19
EU-west
96 ms
0.24%
0.33 USD
627 req/s
32%
18
yes
firewall-20
US-east
149 ms
0.63%
0.264 USD
475 req/s
32%
14
yes
monitor-21
AP-south
85 ms
0.6%
0.166 USD
338 req/s
26%
8
yes
panel-22
EU-west
138 ms
0.48%
0.376 USD
738 req/s
30%
18
yes
program-23
US-east
138 ms
0.61%
0.407 USD
782 req/s
30%
20
yes
circuit-24
AP-south
153 ms
0.72%
0.245 USD
449 req/s
35%
12
yes
pixel-25
EU-west
116 ms
0.39%
0.127 USD
244 req/s
22%
6
yes
monitor-26
US-east
157 ms
1.03%
0.198 USD
397 req/s
5%
8
yes
card-27
AP-south
85 ms
0.1%
0.04 USD
79 req/s
31%
2
yes
firewall-28
EU-west
120 ms
0.36%
0.195 USD
364 req/s
43%
10
yes
Every axis with the range it is scaled to. Each axis is normalised to its own range, which is what lets different units share one plot; two heights on two different axes are not comparable.
Axis
Position
Range
Domain
Direction
No data
Outside range
p95 latency
1 of 6
85 ms to 167 ms
fitted to this data
largest at the bottom
0
0
Error rate
2 of 6
0.1% to 1.03%
fitted to this data
largest at the bottom
0
0
Cost / 1k req
3 of 6
0.04 USD to 0.426 USD
fitted to this data
largest at the bottom
0
0
Throughput
4 of 6
55 req/s to 786 req/s
fitted to this data
largest at the top
0
0
CPU headroom
5 of 6
0% to 100%
fixed by the contract
largest at the top
0
0
Replicas
6 of 6
2 to 22
fitted to this data
largest at the top
0
0
Relationship between each neighbouring pair of axes over the 28 shown records, as a Spearman rank correlation. Only neighbours are listed because only neighbours are joined by a segment; reorder the axes to compare a different pair.
Pair
Rank correlation
Records
What the lines do
p95 latency and Error rate
+0.84
28
segments run mostly parallel
Error rate and Cost / 1k req
+0.13
28
segments are mixed
Cost / 1k req and Throughput
+0.99
28
segments cross heavily
Throughput and CPU headroom
-0.00
28
segments are mixed
CPU headroom and Replicas
+0.18
28
segments are mixed
status="empty" · the axes arrived but not a single record — kept distinct from "loading" and "the fetch failed".
Service fleet · last 24 h
Six mixed-unit axes · latency, error rate and cost are drawn upside down so high is always good
No records to compare yet
The axes are ready; a line appears here for each record that arrives.
status="loading" · the skeleton is only axes and a few polylines, aria-hidden, and does not depend on a timer.
Service fleet · last 24 h
Six mixed-unit axes · latency, error rate and cost are drawn upside down so high is always good
Loading chart
status="error" · the fetch failed; "Try again" only appears when onRetry is passed, and pressing it switches to ready.
Service fleet · last 24 h
Six mixed-unit axes · latency, error rate and cost are drawn upside down so high is always good
Couldn't load this chart
The data source didn't respond.
180 records · the unselected bundle drops to a low opacity scaled by density, so what you read is the shape of the bundle rather than any one line. Drag vertically on any axis to brush; filtered-out lines only fade, they never disappear — that is what makes "which group got filtered out" visible. onSelectionChange is wired: 180 currently match.
Service fleet · last 24 h
180 services · point at the bundle, or brush an axis to pull a subset out of it
EU-west
US-east
AP-south
180 of 180 shown
Point at a line to read it, or Tab into the plot and use the arrow keys. Drag on an axis to filter, drag a header to reorder.
Strongest neighbouring pair: Cost / 1k req and Throughput, rank correlation +1.00 over 180 shown records — their segments cross heavily.
Service fleet · last 24 h: every record with its value on every axis, in the current axis order. No filter is active, so every record counts as shown.
Record
Group
p95 latency (ms)
Error rate (%)
Cost / 1k req (USD)
Throughput (req/s)
CPU headroom (%)
Replicas
In filter
sensor-1
EU-west
99 ms
0.1%
0.118 USD
231 req/s
39%
6
yes
transmitter-2
US-east
157 ms
0.95%
0.233 USD
387 req/s
45%
13
yes
capacitor-3
AP-south
112 ms
0.21%
0.192 USD
365 req/s
40%
10
yes
hard drive-4
EU-west
121 ms
0.58%
0.093 USD
149 req/s
56%
5
yes
bus-5
US-east
124 ms
0.67%
0.141 USD
214 req/s
55%
9
yes
array-6
AP-south
171 ms
0.85%
0.363 USD
696 req/s
37%
17
yes
card-7
EU-west
115 ms
0.27%
0.231 USD
440 req/s
37%
12
yes
application-8
US-east
168 ms
0.74%
0.341 USD
636 req/s
32%
17
yes
card-9
AP-south
124 ms
0.48%
0.31 USD
588 req/s
35%
16
yes
capacitor-10
EU-west
147 ms
0.83%
0.137 USD
222 req/s
45%
7
yes
card-11
US-east
96 ms
0.14%
0.4 USD
799 req/s
27%
20
yes
program-12
AP-south
117 ms
0.47%
0.357 USD
669 req/s
46%
19
yes
capacitor-13
EU-west
119 ms
0.5%
0.094 USD
155 req/s
53%
5
yes
matrix-14
US-east
128 ms
0.8%
0.176 USD
349 req/s
20%
8
yes
driver-15
AP-south
146 ms
0.63%
0.085 USD
142 req/s
33%
4
yes
system-16
EU-west
162 ms
0.91%
0.123 USD
192 req/s
48%
6
yes
protocol-17
US-east
116 ms
0.51%
0.086 USD
148 req/s
28%
4
yes
program-18
AP-south
154 ms
0.65%
0.278 USD
533 req/s
31%
13
yes
port-19
EU-west
118 ms
0.43%
0.175 USD
302 req/s
54%
10
yes
matrix-20
US-east
176 ms
1.25%
0.083 USD
141 req/s
24%
3
yes
matrix-21
AP-south
149 ms
0.82%
0.228 USD
399 req/s
47%
12
yes
pixel-22
EU-west
145 ms
0.57%
0.231 USD
440 req/s
34%
11
yes
firewall-23
US-east
111 ms
0.48%
0.124 USD
188 req/s
64%
8
yes
bandwidth-24
AP-south
132 ms
0.65%
0.215 USD
406 req/s
26%
10
yes
pixel-25
EU-west
119 ms
0.47%
0.405 USD
768 req/s
37%
21
yes
matrix-26
US-east
119 ms
0.29%
0.144 USD
257 req/s
48%
8
yes
interface-27
AP-south
159 ms
0.82%
0.424 USD
802 req/s
30%
21
yes
system-28
EU-west
109 ms
0.3%
0.244 USD
442 req/s
51%
14
yes
array-29
US-east
99 ms
0.09%
0.409 USD
790 req/s
28%
22
yes
pixel-30
AP-south
131 ms
0.64%
0.124 USD
196 req/s
41%
7
yes
capacitor-31
EU-west
159 ms
1.09%
0.079 USD
141 req/s
20%
3
yes
microchip-32
US-east
113 ms
0.34%
0.096 USD
154 req/s
46%
6
yes
feed-33
AP-south
96 ms
0.33%
0.428 USD
831 req/s
43%
22
yes
card-34
EU-west
163 ms
0.91%
0.247 USD
446 req/s
27%
12
yes
matrix-35
US-east
93 ms
0.21%
0.35 USD
684 req/s
28%
18
yes
microchip-36
AP-south
85 ms
0.12%
0.205 USD
392 req/s
28%
11
yes
port-37
EU-west
97 ms
0.28%
0.431 USD
809 req/s
42%
24
yes
panel-38
US-east
134 ms
0.71%
0.127 USD
203 req/s
54%
7
yes
application-39
AP-south
93 ms
0.3%
0.083 USD
118 req/s
56%
6
yes
driver-40
EU-west
119 ms
0.37%
0.171 USD
296 req/s
52%
10
yes
pixel-41
US-east
92 ms
0.27%
0.262 USD
500 req/s
31%
14
yes
transmitter-42
AP-south
145 ms
0.63%
0.073 USD
119 req/s
36%
3
yes
firewall-43
EU-west
105 ms
0.31%
0.242 USD
467 req/s
23%
12
yes
panel-44
US-east
140 ms
0.8%
0.088 USD
118 req/s
61%
5
yes
system-45
AP-south
170 ms
0.8%
0.281 USD
533 req/s
37%
13
yes
circuit-46
EU-west
136 ms
0.61%
0.416 USD
796 req/s
29%
21
yes
circuit-47
US-east
138 ms
0.76%
0.128 USD
203 req/s
56%
7
yes
microchip-48
AP-south
127 ms
0.55%
0.422 USD
815 req/s
30%
21
yes
driver-49
EU-west
105 ms
0.37%
0.182 USD
332 req/s
50%
10
yes
driver-50
US-east
131 ms
0.65%
0.09 USD
152 req/s
25%
4
yes
driver-51
AP-south
166 ms
0.89%
0.26 USD
481 req/s
27%
12
yes
system-52
EU-west
130 ms
0.71%
0.153 USD
261 req/s
47%
8
yes
microchip-53
US-east
90 ms
0.64%
0.205 USD
428 req/s
18%
9
yes
firewall-54
AP-south
103 ms
0.27%
0.37 USD
721 req/s
34%
19
yes
interface-55
EU-west
163 ms
0.72%
0.33 USD
589 req/s
36%
18
yes
hard drive-56
US-east
159 ms
0.86%
0.348 USD
671 req/s
28%
16
yes
protocol-57
AP-south
123 ms
0.51%
0.337 USD
664 req/s
31%
16
yes
card-58
EU-west
127 ms
0.38%
0.356 USD
689 req/s
32%
18
yes
program-59
US-east
127 ms
0.51%
0.107 USD
189 req/s
43%
5
yes
pixel-60
AP-south
86 ms
0.1%
0.371 USD
716 req/s
29%
20
yes
driver-61
EU-west
134 ms
0.69%
0.209 USD
352 req/s
43%
12
yes
program-62
US-east
128 ms
0.41%
0.241 USD
442 req/s
45%
13
yes
monitor-63
AP-south
138 ms
0.6%
0.178 USD
305 req/s
46%
10
yes
card-64
EU-west
133 ms
0.62%
0.228 USD
414 req/s
39%
12
yes
bandwidth-65
US-east
138 ms
0.55%
0.309 USD
591 req/s
27%
15
yes
alarm-66
AP-south
173 ms
0.87%
0.237 USD
414 req/s
47%
12
yes
firewall-67
EU-west
159 ms
0.87%
0.305 USD
558 req/s
28%
15
yes
feed-68
US-east
127 ms
0.4%
0.314 USD
620 req/s
25%
15
yes
hard drive-69
AP-south
156 ms
0.78%
0.078 USD
124 req/s
30%
3
yes
program-70
EU-west
89 ms
0.11%
0.355 USD
702 req/s
28%
18
yes
bus-71
US-east
117 ms
0.33%
0.392 USD
757 req/s
37%
20
yes
sensor-72
AP-south
138 ms
0.66%
0.075 USD
103 req/s
46%
4
yes
matrix-73
EU-west
136 ms
0.54%
0.424 USD
789 req/s
33%
23
yes
bandwidth-74
US-east
73 ms
0.07%
0.049 USD
89 req/s
47%
3
yes
application-75
AP-south
161 ms
0.84%
0.153 USD
252 req/s
42%
8
yes
driver-76
EU-west
185 ms
1%
0.374 USD
703 req/s
33%
18
yes
monitor-77
US-east
153 ms
0.79%
0.236 USD
421 req/s
30%
12
yes
monitor-78
AP-south
162 ms
0.89%
0.078 USD
122 req/s
36%
3
yes
bandwidth-79
EU-west
158 ms
0.76%
0.258 USD
472 req/s
30%
13
yes
bandwidth-80
US-east
94 ms
0.24%
0.17 USD
301 req/s
42%
10
yes
circuit-81
AP-south
79 ms
0.07%
0.155 USD
288 req/s
47%
9
yes
feed-82
EU-west
80 ms
0.22%
0.136 USD
246 req/s
37%
8
yes
transmitter-83
US-east
153 ms
0.93%
0.107 USD
165 req/s
33%
5
yes
bandwidth-84
AP-south
129 ms
0.57%
0.394 USD
768 req/s
22%
19
yes
interface-85
EU-west
109 ms
0.56%
0.079 USD
92 req/s
70%
6
yes
program-86
US-east
137 ms
0.49%
0.314 USD
612 req/s
37%
15
yes
port-87
AP-south
144 ms
0.73%
0.345 USD
655 req/s
31%
17
yes
microchip-88
EU-west
105 ms
0.21%
0.425 USD
814 req/s
38%
23
yes
card-89
US-east
120 ms
0.47%
0.38 USD
716 req/s
28%
20
yes
port-90
AP-south
141 ms
1%
0.081 USD
145 req/s
13%
3
yes
port-91
EU-west
130 ms
0.49%
0.156 USD
281 req/s
31%
8
yes
protocol-92
US-east
90 ms
0.13%
0.033 USD
58 req/s
54%
2
yes
bus-93
AP-south
171 ms
0.87%
0.387 USD
740 req/s
25%
18
yes
hard drive-94
EU-west
144 ms
0.69%
0.173 USD
318 req/s
29%
8
yes
program-95
US-east
94 ms
0.18%
0.384 USD
763 req/s
21%
19
yes
feed-96
AP-south
113 ms
0.21%
0.078 USD
142 req/s
40%
4
yes
driver-97
EU-west
105 ms
0.72%
0.211 USD
421 req/s
32%
10
yes
program-98
US-east
114 ms
0.47%
0.044 USD
65 req/s
43%
2
yes
monitor-99
AP-south
156 ms
1.09%
0.142 USD
265 req/s
27%
6
yes
panel-100
EU-west
107 ms
0.3%
0.398 USD
743 req/s
39%
22
yes
bandwidth-101
US-east
114 ms
0.21%
0.409 USD
797 req/s
26%
21
yes
port-102
AP-south
149 ms
0.89%
0.079 USD
103 req/s
47%
4
yes
application-103
EU-west
122 ms
0.53%
0.419 USD
828 req/s
28%
20
yes
capacitor-104
US-east
152 ms
0.94%
0.057 USD
99 req/s
16%
2
yes
panel-105
AP-south
125 ms
0.37%
0.391 USD
770 req/s
38%
19
yes
application-106
EU-west
101 ms
0.21%
0.372 USD
747 req/s
19%
18
yes
firewall-107
US-east
166 ms
1.19%
0.41 USD
799 req/s
26%
19
yes
bandwidth-108
AP-south
99 ms
0.32%
0.132 USD
239 req/s
39%
7
yes
firewall-109
EU-west
152 ms
0.66%
0.154 USD
281 req/s
25%
7
yes
bandwidth-110
US-east
178 ms
1.31%
0.238 USD
460 req/s
11%
10
yes
panel-111
AP-south
101 ms
0.36%
0.103 USD
173 req/s
58%
6
yes
panel-112
EU-west
138 ms
0.62%
0.422 USD
830 req/s
23%
20
yes
panel-113
US-east
122 ms
0.38%
0.218 USD
415 req/s
28%
11
yes
sensor-114
AP-south
143 ms
0.67%
0.183 USD
333 req/s
34%
9
yes
sensor-115
EU-west
138 ms
0.89%
0.33 USD
659 req/s
24%
15
yes
feed-116
US-east
140 ms
0.69%
0.145 USD
239 req/s
52%
8
yes
bandwidth-117
AP-south
119 ms
0.49%
0.343 USD
625 req/s
36%
19
yes
circuit-118
EU-west
150 ms
0.78%
0.205 USD
352 req/s
52%
11
yes
sensor-119
US-east
77 ms
0.16%
0.149 USD
269 req/s
46%
9
yes
panel-120
AP-south
134 ms
0.5%
0.405 USD
789 req/s
22%
20
yes
feed-121
EU-west
106 ms
0.38%
0.337 USD
637 req/s
33%
18
yes
alarm-122
US-east
141 ms
0.72%
0.206 USD
357 req/s
44%
11
yes
matrix-123
AP-south
127 ms
0.43%
0.294 USD
559 req/s
41%
15
yes
circuit-124
EU-west
133 ms
0.53%
0.204 USD
382 req/s
34%
10
yes
array-125
US-east
104 ms
0.37%
0.164 USD
281 req/s
56%
10
yes
circuit-126
AP-south
157 ms
0.82%
0.177 USD
284 req/s
45%
10
yes
microchip-127
EU-west
120 ms
0.31%
0.395 USD
752 req/s
36%
21
yes
transmitter-128
US-east
122 ms
0.52%
0.316 USD
618 req/s
36%
15
yes
capacitor-129
AP-south
165 ms
0.74%
0.377 USD
707 req/s
30%
19
yes
interface-130
EU-west
121 ms
0.51%
0.177 USD
327 req/s
37%
9
yes
pixel-131
US-east
113 ms
0.63%
0.1 USD
150 req/s
58%
6
yes
sensor-132
AP-south
102 ms
0.32%
0.075 USD
130 req/s
51%
4
yes
port-133
EU-west
141 ms
1.13%
0.056 USD
86 req/s
25%
2
yes
protocol-134
US-east
114 ms
0.35%
0.239 USD
423 req/s
46%
14
yes
protocol-135
AP-south
118 ms
0.39%
0.294 USD
547 req/s
49%
16
yes
driver-136
EU-west
147 ms
0.79%
0.28 USD
531 req/s
34%
13
yes
bandwidth-137
US-east
114 ms
0.21%
0.347 USD
692 req/s
38%
17
yes
circuit-138
AP-south
80 ms
0.18%
0.254 USD
482 req/s
44%
14
yes
array-139
EU-west
99 ms
0.22%
0.189 USD
357 req/s
31%
10
yes
interface-140
US-east
157 ms
0.88%
0.17 USD
277 req/s
39%
9
yes
firewall-141
AP-south
138 ms
0.74%
0.276 USD
505 req/s
32%
14
yes
alarm-142
EU-west
80 ms
0.15%
0.109 USD
171 req/s
59%
8
yes
microchip-143
US-east
149 ms
0.78%
0.446 USD
830 req/s
39%
23
yes
bus-144
AP-south
106 ms
0.4%
0.068 USD
88 req/s
68%
5
yes
firewall-145
EU-west
81 ms
0.27%
0.059 USD
85 req/s
57%
4
yes
card-146
US-east
156 ms
0.82%
0.372 USD
686 req/s
40%
19
yes
interface-147
AP-south
164 ms
1.19%
0.11 USD
199 req/s
5%
4
yes
driver-148
EU-west
110 ms
0.64%
0.089 USD
116 req/s
74%
6
yes
feed-149
US-east
137 ms
0.91%
0.159 USD
308 req/s
18%
7
yes
application-150
AP-south
90 ms
0.19%
0.153 USD
294 req/s
34%
8
yes
program-151
EU-west
102 ms
0.41%
0.059 USD
60 req/s
86%
5
yes
microchip-152
US-east
101 ms
0.65%
0.114 USD
232 req/s
21%
5
yes
sensor-153
AP-south
107 ms
0.33%
0.225 USD
435 req/s
25%
11
yes
matrix-154
EU-west
144 ms
0.76%
0.092 USD
151 req/s
34%
4
yes
circuit-155
US-east
99 ms
0.66%
0.358 USD
720 req/s
29%
17
yes
capacitor-156
AP-south
83 ms
0.09%
0.07 USD
113 req/s
51%
5
yes
transmitter-157
EU-west
143 ms
0.73%
0.12 USD
201 req/s
44%
6
yes
alarm-158
US-east
126 ms
0.5%
0.236 USD
459 req/s
20%
11
yes
card-159
AP-south
103 ms
0.18%
0.413 USD
813 req/s
29%
21
yes
firewall-160
EU-west
142 ms
0.73%
0.43 USD
827 req/s
34%
21
yes
interface-161
US-east
86 ms
0.24%
0.155 USD
278 req/s
41%
9
yes
pixel-162
AP-south
120 ms
0.36%
0.244 USD
471 req/s
34%
12
yes
program-163
EU-west
155 ms
0.85%
0.179 USD
303 req/s
41%
9
yes
program-164
US-east
120 ms
0.84%
0.221 USD
444 req/s
16%
10
yes
feed-165
AP-south
131 ms
0.63%
0.186 USD
325 req/s
40%
10
yes
bus-166
EU-west
140 ms
0.75%
0.314 USD
561 req/s
36%
17
yes
interface-167
US-east
126 ms
0.57%
0.448 USD
851 req/s
30%
23
yes
card-168
AP-south
53 ms
0.06%
0.059 USD
83 req/s
73%
5
yes
driver-169
EU-west
136 ms
0.54%
0.254 USD
474 req/s
32%
13
yes
port-170
US-east
180 ms
0.96%
0.409 USD
787 req/s
27%
19
yes
matrix-171
AP-south
116 ms
0.45%
0.337 USD
646 req/s
42%
17
yes
matrix-172
EU-west
154 ms
1%
0.292 USD
580 req/s
13%
13
yes
card-173
US-east
137 ms
0.6%
0.35 USD
658 req/s
40%
18
yes
pixel-174
AP-south
155 ms
1.09%
0.079 USD
132 req/s
15%
3
yes
interface-175
EU-west
168 ms
0.95%
0.346 USD
623 req/s
45%
18
yes
monitor-176
US-east
155 ms
0.96%
0.204 USD
338 req/s
42%
11
yes
capacitor-177
AP-south
152 ms
0.62%
0.282 USD
509 req/s
32%
15
yes
alarm-178
EU-west
93 ms
0.14%
0.205 USD
375 req/s
38%
12
yes
system-179
US-east
133 ms
0.48%
0.397 USD
779 req/s
27%
19
yes
alarm-180
AP-south
115 ms
0.4%
0.063 USD
88 req/s
53%
4
yes
Every axis with the range it is scaled to. Each axis is normalised to its own range, which is what lets different units share one plot; two heights on two different axes are not comparable.
Axis
Position
Range
Domain
Direction
No data
Outside range
p95 latency
1 of 6
53 ms to 185 ms
fitted to this data
largest at the bottom
0
0
Error rate
2 of 6
0.06% to 1.31%
fitted to this data
largest at the bottom
0
0
Cost / 1k req
3 of 6
0.033 USD to 0.448 USD
fitted to this data
largest at the bottom
0
0
Throughput
4 of 6
58 req/s to 851 req/s
fitted to this data
largest at the top
0
0
CPU headroom
5 of 6
0% to 100%
fixed by the contract
largest at the top
0
0
Replicas
6 of 6
2 to 24
fitted to this data
largest at the top
0
0
Relationship between each neighbouring pair of axes over the 180 shown records, as a Spearman rank correlation. Only neighbours are listed because only neighbours are joined by a segment; reorder the axes to compare a different pair.
Pair
Rank correlation
Records
What the lines do
p95 latency and Error rate
+0.89
180
segments run mostly parallel
Error rate and Cost / 1k req
-0.07
180
segments are mixed
Cost / 1k req and Throughput
+1.00
180
segments cross heavily
Throughput and CPU headroom
-0.43
180
segments cross heavily
CPU headroom and Replicas
-0.27
180
segments are mixed
Every degenerate case on one card: priority is 3 on every record (a flat axis, drawn down the middle and labelled all equal) · drift goes negative · score has a fixed 0–10 domain but one record is 12.4 (clamped to the ceiling and flagged, still compared at its true value when brushed) · one record has no runtime (the polyline breaks at that axis instead of dropping to 0) · one label is wider than its lane (truncated by the browser, in full in the readout and the table).
Batch jobs · edge cases in one plot
Flat axis · negative values · a value past a fixed domain · a missing value · a label that cannot fit
batch
ml
compliance
4 of 4 shown
Point at a line to read it, or Tab into the plot and use the arrow keys. Drag on an axis to filter, drag a header to reorder.
Strongest neighbouring pair: Quality score (fixed 0–10 scale) and Runtime, rank correlation -0.50 over 3 shown records — their segments cross heavily.
Batch jobs · edge cases in one plot: every record with its value on every axis, in the current axis order. No filter is active, so every record counts as shown.
Every axis with the range it is scaled to. Each axis is normalised to its own range, which is what lets different units share one plot; two heights on two different axes are not comparable.
Axis
Position
Range
Domain
Direction
No data
Outside range
Priority
1 of 4
3 on every record
fitted to this data
largest at the top
0
0
Budget drift
2 of 4
-18.5% to 240.8%
fitted to this data
largest at the top
0
0
Quality score (fixed 0–10 scale)
3 of 4
0 to 10
fixed by the contract
largest at the top
0
1
Runtime
4 of 4
3 min to 42 min
fitted to this data
largest at the top
1
0
Relationship between each neighbouring pair of axes over the 4 shown records, as a Spearman rank correlation. Only neighbours are listed because only neighbours are joined by a segment; reorder the axes to compare a different pair.
Pair
Rank correlation
Records
What the lines do
Priority and Budget drift
no reading: fewer than three records, or one side all equal
4
no reading
Budget drift and Quality score (fixed 0–10 scale)
-0.20
4
segments are mixed
Quality score (fixed 0–10 scale) and Runtime
-0.50
3
segments cross heavily
A single record: every fitted axis now has min === max and is drawn down its middle; CPU headroom, the one axis with a fixed domain, still places it at a real height.
Service fleet · one service reporting
One record: every fitted axis collapses; the fixed 0–100 axis still places it
EU-west
1 of 1 shown
Point at a line to read it, or Tab into the plot and use the arrow keys. Drag on an axis to filter, drag a header to reorder.
No neighbouring pair has enough varying records to report a relationship yet.
Service fleet · one service reporting: every record with its value on every axis, in the current axis order. No filter is active, so every record counts as shown.
Record
Group
p95 latency (ms)
Error rate (%)
Cost / 1k req (USD)
Throughput (req/s)
CPU headroom (%)
Replicas
In filter
application-1
EU-west
114 ms
0.28%
0.076 USD
137 req/s
40%
4
yes
Every axis with the range it is scaled to. Each axis is normalised to its own range, which is what lets different units share one plot; two heights on two different axes are not comparable.
Axis
Position
Range
Domain
Direction
No data
Outside range
p95 latency
1 of 6
114 ms on every record
fitted to this data
largest at the bottom
0
0
Error rate
2 of 6
0.28% on every record
fitted to this data
largest at the bottom
0
0
Cost / 1k req
3 of 6
0.076 USD on every record
fitted to this data
largest at the bottom
0
0
Throughput
4 of 6
137 req/s on every record
fitted to this data
largest at the top
0
0
CPU headroom
5 of 6
0% to 100%
fixed by the contract
largest at the top
0
0
Replicas
6 of 6
4 on every record
fitted to this data
largest at the top
0
0
Relationship between each neighbouring pair of axes over the 1 shown records, as a Spearman rank correlation. Only neighbours are listed because only neighbours are joined by a segment; reorder the axes to compare a different pair.
Pair
Rank correlation
Records
What the lines do
p95 latency and Error rate
no reading: fewer than three records, or one side all equal
1
no reading
Error rate and Cost / 1k req
no reading: fewer than three records, or one side all equal
1
no reading
Cost / 1k req and Throughput
no reading: fewer than three records, or one side all equal
1
no reading
Throughput and CPU headroom
no reading: fewer than three records, or one side all equal
1
no reading
CPU headroom and Replicas
no reading: fewer than three records, or one side all equal
1
no reading
Two axes sharing one key · nothing crashes if it is drawn, but the second axis silently plots the first one's numbers and every reading taken off the picture is wrong — so the contract refuses it and names the offending key.
Vendor shortlist · unusable axes
Two axes were given the same key
This comparison can't be drawn
Two axes share the key "score"; the key is how a record finds its value, so one of them would draw the other's numbers.
brushable={false} + reorderable={false} + bandHeight={180} · the compact embedded version: no brush handles, no axis reordering, but axis headers still flip direction on click (or Enter), and hover readouts, keyboard traversal and the sr-only table all keep working.
Service fleet · last 24 h
Six mixed-unit axes · latency, error rate and cost are drawn upside down so high is always good
EU-west
US-east
AP-south
28 of 28 shown
Point at a line to read it, or Tab into the plot and use the arrow keys. Drag on an axis to filter, drag a header to reorder.
Strongest neighbouring pair: Cost / 1k req and Throughput, rank correlation +0.99 over 28 shown records — their segments cross heavily.
Service fleet · last 24 h: every record with its value on every axis, in the current axis order. No filter is active, so every record counts as shown.
Record
Group
p95 latency (ms)
Error rate (%)
Cost / 1k req (USD)
Throughput (req/s)
CPU headroom (%)
Replicas
In filter
pixel-1
EU-west
90 ms
0.19%
0.351 USD
669 req/s
30%
19
yes
driver-2
US-east
126 ms
0.44%
0.337 USD
661 req/s
20%
16
yes
sensor-3
AP-south
108 ms
0.53%
0.315 USD
640 req/s
20%
15
yes
port-4
EU-west
113 ms
0.72%
0.114 USD
229 req/s
19%
5
yes
capacitor-5
US-east
105 ms
0.37%
0.283 USD
515 req/s
43%
16
yes
driver-6
AP-south
146 ms
0.8%
0.378 USD
727 req/s
35%
18
yes
program-7
EU-west
164 ms
0.76%
0.314 USD
559 req/s
45%
17
yes
sensor-8
US-east
131 ms
0.66%
0.223 USD
412 req/s
39%
11
yes
card-9
AP-south
164 ms
0.87%
0.214 USD
390 req/s
26%
10
yes
monitor-10
EU-west
149 ms
0.86%
0.137 USD
204 req/s
55%
8
yes
microchip-11
US-east
109 ms
0.66%
0.222 USD
458 req/s
23%
10
yes
port-12
AP-south
167 ms
0.86%
0.426 USD
786 req/s
44%
22
yes
interface-13
EU-west
96 ms
0.32%
0.162 USD
295 req/s
38%
9
yes
port-14
US-east
149 ms
0.85%
0.401 USD
774 req/s
29%
19
yes
monitor-15
AP-south
146 ms
0.86%
0.394 USD
735 req/s
39%
20
yes
bus-16
EU-west
103 ms
0.41%
0.046 USD
55 req/s
74%
3
yes
application-17
US-east
122 ms
0.42%
0.401 USD
768 req/s
39%
21
yes
firewall-18
AP-south
87 ms
0.11%
0.394 USD
764 req/s
31%
21
yes
card-19
EU-west
96 ms
0.24%
0.33 USD
627 req/s
32%
18
yes
firewall-20
US-east
149 ms
0.63%
0.264 USD
475 req/s
32%
14
yes
monitor-21
AP-south
85 ms
0.6%
0.166 USD
338 req/s
26%
8
yes
panel-22
EU-west
138 ms
0.48%
0.376 USD
738 req/s
30%
18
yes
program-23
US-east
138 ms
0.61%
0.407 USD
782 req/s
30%
20
yes
circuit-24
AP-south
153 ms
0.72%
0.245 USD
449 req/s
35%
12
yes
pixel-25
EU-west
116 ms
0.39%
0.127 USD
244 req/s
22%
6
yes
monitor-26
US-east
157 ms
1.03%
0.198 USD
397 req/s
5%
8
yes
card-27
AP-south
85 ms
0.1%
0.04 USD
79 req/s
31%
2
yes
firewall-28
EU-west
120 ms
0.36%
0.195 USD
364 req/s
43%
10
yes
Every axis with the range it is scaled to. Each axis is normalised to its own range, which is what lets different units share one plot; two heights on two different axes are not comparable.
Axis
Position
Range
Domain
Direction
No data
Outside range
p95 latency
1 of 6
85 ms to 167 ms
fitted to this data
largest at the bottom
0
0
Error rate
2 of 6
0.1% to 1.03%
fitted to this data
largest at the bottom
0
0
Cost / 1k req
3 of 6
0.04 USD to 0.426 USD
fitted to this data
largest at the bottom
0
0
Throughput
4 of 6
55 req/s to 786 req/s
fitted to this data
largest at the top
0
0
CPU headroom
5 of 6
0% to 100%
fixed by the contract
largest at the top
0
0
Replicas
6 of 6
2 to 22
fitted to this data
largest at the top
0
0
Relationship between each neighbouring pair of axes over the 28 shown records, as a Spearman rank correlation. Only neighbours are listed because only neighbours are joined by a segment; reorder the axes to compare a different pair.
Zyeon UI