This repository has no description
configs.gg23.internet: monitor the Vodafone Station, latency and speedtests, replace HA SpeedTest
author
lassulus
committer
lassulus
date
(Oct 18, 2026, 10:33 AM CEST)
commit
83397339
833973391e8facacf51e7848e3f636b4fe87a621
parent
498e5909
498e59090ba535c8079707dea39fbbe773474020
+928
+392
configs/gg23/internet-hass.nix
+392
configs/gg23/internet-hass.nix
···
1
+
# Home Assistant side of ./internet.nix: sensors that query VictoriaMetrics on
2
+
# styx (the gg23 monitoring server) and an "Internet" dashboard explaining
3
+
# why the internet is slow. Both are mounted into the container as a package
4
+
# and a YAML dashboard (configs/hass loads /config/packages); generated as
5
+
# JSON, which HA's YAML loader reads as well.
6
+
{
7
+
lib,
8
+
pkgs,
9
+
self,
10
+
...
11
+
}:
12
+
let
13
+
inherit (import ../monitoring/default.nix) ports;
14
+
# IPv4: VictoriaMetrics doesn't listen on IPv6, styx.r resolves to it first.
15
+
vm = "http://${self.inputs.kartei.hosts.styx.nets.retiolum.ip4.addr}:${toString ports.victoriametrics}/api/v1/query";
16
+
17
+
# booked Vodafone cable rate in Mbit/s (speedtests from acheron: ~920/50)
18
+
line = {
19
+
down = 1000;
20
+
up = 50;
21
+
};
22
+
23
+
ping = sel: {
24
+
latency = ''avg(rate(smokeping_response_duration_seconds_sum{host="acheron",${sel}}[5m]) / rate(smokeping_response_duration_seconds_count{host="acheron",${sel}}[5m])) * 1000'';
25
+
p95 = ''histogram_quantile(0.95, sum(rate(smokeping_response_duration_seconds_bucket{host="acheron",${sel}}[5m])) by (le)) * 1000'';
26
+
loss = ''100 * (1 - sum(rate(smokeping_response_duration_seconds_count{host="acheron",${sel}}[5m])) / sum(rate(smokeping_requests_total{host="acheron",${sel}}[5m])))'';
27
+
};
28
+
internet = ping ''target_name=~"cloudflare|google"'';
29
+
cmts = ping ''target_name="cmts"'';
30
+
station = ping ''target_name="station"'';
31
+
wan = dir: ''rate(host_network_${dir}_bytes_total{host="acheron",device="et0"}[2m]) * 8 / 1e6'';
32
+
docsis = ''{host="acheron",channel_type="SC-QAM"}'';
33
+
34
+
# id -> sensor.internet_<id>
35
+
sensors = {
36
+
latency = {
37
+
name = "Latency";
38
+
query = internet.latency;
39
+
unit = "ms";
40
+
icon = "mdi:timer-outline";
41
+
};
42
+
latency_p95 = {
43
+
name = "Latency p95";
44
+
query = internet.p95;
45
+
unit = "ms";
46
+
icon = "mdi:timer-alert-outline";
47
+
};
48
+
loss = {
49
+
name = "Packet loss";
50
+
query = internet.loss;
51
+
unit = "%";
52
+
icon = "mdi:lan-disconnect";
53
+
};
54
+
cmts_latency = {
55
+
name = "Cable latency";
56
+
query = cmts.latency;
57
+
unit = "ms";
58
+
icon = "mdi:timer-outline";
59
+
};
60
+
cmts_loss = {
61
+
name = "Cable packet loss";
62
+
query = cmts.loss;
63
+
unit = "%";
64
+
icon = "mdi:lan-disconnect";
65
+
};
66
+
station_latency = {
67
+
name = "Station latency";
68
+
query = station.latency;
69
+
unit = "ms";
70
+
icon = "mdi:router-network";
71
+
};
72
+
station_loss = {
73
+
name = "Station packet loss";
74
+
query = station.loss;
75
+
unit = "%";
76
+
icon = "mdi:router-network";
77
+
};
78
+
download = {
79
+
name = "Download";
80
+
query = wan "receive";
81
+
unit = "Mbit/s";
82
+
device_class = "data_rate";
83
+
icon = "mdi:download";
84
+
};
85
+
upload = {
86
+
name = "Upload";
87
+
query = wan "transmit";
88
+
unit = "Mbit/s";
89
+
device_class = "data_rate";
90
+
icon = "mdi:upload";
91
+
};
92
+
# hourly, ./internet-speedtest.nix; last result within 2h
93
+
speedtest_download = {
94
+
name = "Speedtest download";
95
+
query = ''last_over_time(speedtest_download_bits_per_second{host="acheron"}[2h]) / 1e6'';
96
+
unit = "Mbit/s";
97
+
device_class = "data_rate";
98
+
icon = "mdi:speedometer";
99
+
};
100
+
speedtest_upload = {
101
+
name = "Speedtest upload";
102
+
query = ''last_over_time(speedtest_upload_bits_per_second{host="acheron"}[2h]) / 1e6'';
103
+
unit = "Mbit/s";
104
+
device_class = "data_rate";
105
+
icon = "mdi:speedometer";
106
+
};
107
+
speedtest_latency = {
108
+
name = "Speedtest latency";
109
+
query = ''last_over_time(speedtest_latency_seconds{host="acheron"}[2h]) * 1000'';
110
+
unit = "ms";
111
+
icon = "mdi:timer-outline";
112
+
};
113
+
ds_snr_min = {
114
+
name = "Downstream SNR min";
115
+
query = ''min(vodafone_station_docsis_snr_db{direction="downstream",host="acheron"})'';
116
+
unit = "dB";
117
+
icon = "mdi:signal";
118
+
};
119
+
ds_power_min = {
120
+
name = "Downstream power min";
121
+
query = ''min(vodafone_station_docsis_power_dbmv{direction="downstream",${lib.removePrefix "{" docsis})'';
122
+
unit = "dBmV";
123
+
icon = "mdi:arrow-down-bold";
124
+
};
125
+
ds_power_max = {
126
+
name = "Downstream power max";
127
+
query = ''max(vodafone_station_docsis_power_dbmv{direction="downstream",${lib.removePrefix "{" docsis})'';
128
+
unit = "dBmV";
129
+
icon = "mdi:arrow-down-bold";
130
+
};
131
+
us_power_max = {
132
+
name = "Upstream power max";
133
+
query = ''max(vodafone_station_docsis_power_dbmv{direction="upstream",${lib.removePrefix "{" docsis})'';
134
+
unit = "dBmV";
135
+
icon = "mdi:arrow-up-bold";
136
+
};
137
+
unlocked = {
138
+
name = "Unlocked channels";
139
+
query = ''count(vodafone_station_docsis_locked{host="acheron"} == 0) or vector(0)'';
140
+
icon = "mdi:lock-open-alert";
141
+
};
142
+
docsis_online = {
143
+
name = "DOCSIS online";
144
+
query = ''max(vodafone_station_docsis_online{host="acheron"})'';
145
+
icon = "mdi:cable-data";
146
+
};
147
+
station_up = {
148
+
name = "Station scrape";
149
+
query = ''max(vodafone_station_up{host="acheron"})'';
150
+
icon = "mdi:router-wireless";
151
+
};
152
+
station_uptime = {
153
+
name = "Station uptime";
154
+
query = ''max(vodafone_station_uptime_seconds{host="acheron"})'';
155
+
unit = "s";
156
+
device_class = "duration";
157
+
icon = "mdi:restart";
158
+
};
159
+
};
160
+
161
+
restEntry = id: s: {
162
+
resource = vm;
163
+
params.query = s.query;
164
+
scan_interval = 60;
165
+
timeout = 20;
166
+
sensor = [
167
+
(
168
+
{
169
+
name = "Internet ${s.name}";
170
+
unique_id = "gg23_internet_${id}";
171
+
# no series yet (e.g. rate() right after a restart): unavailable
172
+
availability = "{{ value_json is defined and value_json.data.result | length > 0 }}";
173
+
value_template = "{{ (value_json.data.result[0].value[1] | float) | round(2) }}";
174
+
state_class = "measurement";
175
+
inherit (s) icon;
176
+
}
177
+
// lib.optionalAttrs (s ? unit) { unit_of_measurement = s.unit; }
178
+
// lib.optionalAttrs (s ? device_class) { inherit (s) device_class; }
179
+
)
180
+
];
181
+
};
182
+
183
+
# HA derives the entity id from the name ("Internet Latency p95" ->
184
+
# sensor.internet_latency_p95); unique_id doesn't influence it.
185
+
e =
186
+
id:
187
+
"sensor.internet_" + lib.toLower (lib.replaceStrings [ " " ] [ "_" ] (sensors.${id}.name or id));
188
+
189
+
# Each check: the problem it names, in the order things are upstream of
190
+
# each other (LAN cable -> Station -> cable segment -> backbone). Thresholds
191
+
# for DOCSIS levels follow the usual Vodafone cable level tables.
192
+
diagnosis = ''
193
+
{%- set n = namespace(p=[]) -%}
194
+
{%- if states('${e "station_up"}') in ['0.0', '0'] -%}
195
+
{%- set n.p = n.p + ['Station web UI not scrapeable'] -%}
196
+
{%- endif -%}
197
+
{%- if states('${e "docsis_online"}') in ['0.0', '0'] -%}
198
+
{%- set n.p = n.p + ['Cable modem not DOCSIS online'] -%}
199
+
{%- endif -%}
200
+
{%- if states('${e "station_loss"}') | float(0) > 1 -%}
201
+
{%- set n.p = n.p + ['Loss acheron->Station ' ~ states('${e "station_loss"}') ~ '% (LAN cable/port)'] -%}
202
+
{%- endif -%}
203
+
{%- if states('${e "unlocked"}') | float(0) > 0 -%}
204
+
{%- set n.p = n.p + [states('${e "unlocked"}') | int ~ ' DOCSIS channels unlocked'] -%}
205
+
{%- endif -%}
206
+
{%- if states('${e "ds_snr_min"}') | float(99) < 33 -%}
207
+
{%- set n.p = n.p + ['Downstream SNR low (' ~ states('${e "ds_snr_min"}') ~ ' dB)'] -%}
208
+
{%- endif -%}
209
+
{%- if states('${e "ds_power_min"}') | float(0) < -4 or states('${e "ds_power_max"}') | float(0) > 13 -%}
210
+
{%- set n.p = n.p + ['Downstream level out of range (' ~ states('${e "ds_power_min"}') ~ '..' ~ states('${e "ds_power_max"}') ~ ' dBmV)'] -%}
211
+
{%- endif -%}
212
+
{%- if states('${e "us_power_max"}') | float(0) > 51 -%}
213
+
{%- set n.p = n.p + ['Upstream level high (' ~ states('${e "us_power_max"}') ~ ' dBmV): return path attenuated/noisy'] -%}
214
+
{%- endif -%}
215
+
{%- if states('${e "cmts_loss"}') | float(0) > 1 -%}
216
+
{%- set n.p = n.p + ['Loss on the cable segment ' ~ states('${e "cmts_loss"}') ~ '%'] -%}
217
+
{%- endif -%}
218
+
{%- if states('${e "cmts_latency"}') | float(0) > 30 -%}
219
+
{%- set n.p = n.p + ['Cable segment slow (' ~ states('${e "cmts_latency"}') ~ ' ms to CMTS): node congested'] -%}
220
+
{%- endif -%}
221
+
{%- if states('${e "loss"}') | float(0) > 1 and states('${e "cmts_loss"}') | float(0) <= 1 -%}
222
+
{%- set n.p = n.p + ['Loss behind the CMTS ' ~ states('${e "loss"}') ~ '% (Vodafone backbone)'] -%}
223
+
{%- endif -%}
224
+
{#- the hourly speedtest saturates the line itself: skip the 5 min window after a result -#}
225
+
{%- set st = states['${e "speedtest_download"}'] -%}
226
+
{%- set after_speedtest = st is not none and (now() - st.last_changed).total_seconds() < 420 -%}
227
+
{%- if states('${e "latency_p95"}') | float(0) > 60 and not after_speedtest -%}
228
+
{%- set n.p = n.p + ['Latency spikes (p95 ' ~ states('${e "latency_p95"}') ~ ' ms) at ' ~ states('${e "download"}') ~ '/' ~ states('${e "upload"}') ~ ' Mbit/s: bufferbloat or congestion'] -%}
229
+
{%- endif -%}
230
+
{%- if states('${e "station_uptime"}') | float(99999) < 3600 -%}
231
+
{%- set n.p = n.p + ['Station rebooted in the last hour'] -%}
232
+
{%- endif -%}
233
+
{%- if states('${e "speedtest_download"}') | float(9999) < 0.3 * ${toString line.down} -%}
234
+
{%- set n.p = n.p + ['Speedtest only ' ~ states('${e "speedtest_download"}') ~ ' of ${toString line.down} Mbit/s down'] -%}
235
+
{%- endif -%}
236
+
{%- if states('${e "speedtest_upload"}') | float(9999) < 0.3 * ${toString line.up} -%}
237
+
{%- set n.p = n.p + ['Speedtest only ' ~ states('${e "speedtest_upload"}') ~ ' of ${toString line.up} Mbit/s up'] -%}
238
+
{%- endif -%}
239
+
'';
240
+
241
+
package = {
242
+
rest = lib.mapAttrsToList restEntry sensors;
243
+
template = [
244
+
{
245
+
sensor = [
246
+
{
247
+
name = "Internet diagnosis";
248
+
unique_id = "gg23_internet_diagnosis";
249
+
icon = "mdi:stethoscope";
250
+
# the problem itself when there is one, else the count
251
+
state = "${diagnosis}{{ 'OK' if not n.p else n.p[0] if n.p | count == 1 else n.p | count ~ ' problems' }}";
252
+
attributes.problems = "${diagnosis}{{ ('- ' ~ n.p | join('\\n- ')) if n.p else 'nothing found' }}";
253
+
}
254
+
];
255
+
}
256
+
];
257
+
lovelace.dashboards.internet-gg23 = {
258
+
mode = "yaml";
259
+
title = "Internet";
260
+
icon = "mdi:wan";
261
+
show_in_sidebar = true;
262
+
filename = "dashboards/internet.yaml";
263
+
};
264
+
};
265
+
266
+
graph = title: hours: ids: {
267
+
type = "history-graph";
268
+
inherit title;
269
+
hours_to_show = hours;
270
+
entities = map e ids;
271
+
};
272
+
273
+
dashboard = {
274
+
title = "Internet";
275
+
views = [
276
+
{
277
+
title = "Internet";
278
+
path = "internet";
279
+
type = "sections";
280
+
max_columns = 3;
281
+
sections = [
282
+
{
283
+
type = "grid";
284
+
cards = [
285
+
{
286
+
type = "markdown";
287
+
title = "Why does it suck?";
288
+
content = ''
289
+
**{{ states('${e "diagnosis"}') }}**
290
+
291
+
{{ state_attr('${e "diagnosis"}', 'problems') }}
292
+
293
+
Path: acheron → Station (LAN) → CMTS (cable) → internet.
294
+
Loss/latency appearing first at a hop points at that segment.
295
+
'';
296
+
}
297
+
{
298
+
type = "entities";
299
+
title = "Now (5 min)";
300
+
entities = map e [
301
+
"download"
302
+
"upload"
303
+
"latency"
304
+
"latency_p95"
305
+
"loss"
306
+
"cmts_latency"
307
+
"cmts_loss"
308
+
"station_latency"
309
+
"station_loss"
310
+
];
311
+
}
312
+
];
313
+
}
314
+
{
315
+
type = "grid";
316
+
cards = [
317
+
(graph "Latency" 24 [
318
+
"latency"
319
+
"latency_p95"
320
+
"cmts_latency"
321
+
"station_latency"
322
+
])
323
+
(graph "Packet loss" 24 [
324
+
"loss"
325
+
"cmts_loss"
326
+
"station_loss"
327
+
])
328
+
(graph "Throughput (WAN)" 24 [
329
+
"download"
330
+
"upload"
331
+
])
332
+
{
333
+
type = "entities";
334
+
title = "Speedtest (hourly, from acheron)";
335
+
entities = map e [
336
+
"speedtest_download"
337
+
"speedtest_upload"
338
+
"speedtest_latency"
339
+
];
340
+
}
341
+
(graph "Speedtest" 168 [
342
+
"speedtest_download"
343
+
"speedtest_upload"
344
+
])
345
+
];
346
+
}
347
+
{
348
+
type = "grid";
349
+
cards = [
350
+
{
351
+
type = "entities";
352
+
title = "Cable modem";
353
+
entities = map e [
354
+
"docsis_online"
355
+
"unlocked"
356
+
"ds_snr_min"
357
+
"ds_power_min"
358
+
"ds_power_max"
359
+
"us_power_max"
360
+
"station_uptime"
361
+
"station_up"
362
+
];
363
+
}
364
+
(graph "DOCSIS levels" 168 [
365
+
"ds_snr_min"
366
+
"ds_power_min"
367
+
"ds_power_max"
368
+
"us_power_max"
369
+
])
370
+
{
371
+
type = "markdown";
372
+
content = ''
373
+
Good ranges: downstream -4..13 dBmV, SNR ≥ 33 dB (256QAM),
374
+
upstream 35..51 dBmV. Per channel and longer history:
375
+
Grafana "Internet" on styx.r:${toString ports.grafana}.
376
+
'';
377
+
}
378
+
];
379
+
}
380
+
];
381
+
}
382
+
];
383
+
};
384
+
385
+
json = name: value: pkgs.writeText name (builtins.toJSON value);
386
+
in
387
+
{
388
+
virtualisation.oci-containers.containers.homeassistant.volumes = [
389
+
"${json "internet.yaml" package}:/config/packages/internet.yaml:ro"
390
+
"${json "internet-dashboard.yaml" dashboard}:/config/dashboards/internet.yaml:ro"
391
+
];
392
+
}
+114
configs/gg23/internet-speedtest.nix
+114
configs/gg23/internet-speedtest.nix
···
1
+
# Hourly speedtest from acheron, replacing Home Assistant's SpeedTest
2
+
# integration: speedtest.net geolocates our Vodafone IP near Sicily, so HA
3
+
# only offered servers in Italy/Tunisia/Algeria, and its single-threaded
4
+
# Python client collapsed on their 500+ ms latency (42 Mbit/s on a 1000/50
5
+
# line). speedtest-go runs parallel connections against a German server.
6
+
#
7
+
# Each run appends one JSON line to results.jsonl (also the history); the
8
+
# Vector agent tails it into VictoriaMetrics. A run moves ~1.4 GB and
9
+
# saturates the line for ~30 s, hence hourly and not more often.
10
+
{
11
+
config,
12
+
lib,
13
+
pkgs,
14
+
...
15
+
}:
16
+
let
17
+
dir = "/var/lib/speedtest";
18
+
in
19
+
{
20
+
# Static user: a DynamicUser's state lives under /var/lib/private (0700),
21
+
# where Vector (itself a DynamicUser) can't read it.
22
+
users.users.speedtest = {
23
+
isSystemUser = true;
24
+
group = "speedtest";
25
+
};
26
+
users.groups.speedtest = { };
27
+
28
+
systemd.services.speedtest = {
29
+
description = "Speedtest against a German speedtest.net server";
30
+
after = [ "network-online.target" ];
31
+
wants = [ "network-online.target" ];
32
+
serviceConfig = {
33
+
Type = "oneshot";
34
+
User = "speedtest";
35
+
Group = "speedtest";
36
+
StateDirectory = "speedtest";
37
+
StateDirectoryMode = "0755";
38
+
UMask = "0022";
39
+
TimeoutStartSec = "5min";
40
+
};
41
+
script = ''
42
+
${lib.getExe pkgs.speedtest-go} --jsonl --filter-cc DE >> ${dir}/results.jsonl
43
+
'';
44
+
};
45
+
46
+
systemd.timers.speedtest = {
47
+
wantedBy = [ "timers.target" ];
48
+
timerConfig = {
49
+
OnCalendar = "hourly";
50
+
RandomizedDelaySec = "10min";
51
+
};
52
+
};
53
+
54
+
monitoring.extraMetrics = [ "speedtest_metrics" ];
55
+
56
+
services.vector.settings = {
57
+
sources.speedtest_raw = {
58
+
type = "file";
59
+
include = [ "${dir}/results.jsonl" ];
60
+
};
61
+
# speedtest-go: speeds in bytes/s, latencies in ns
62
+
transforms.speedtest = {
63
+
type = "remap";
64
+
inputs = [ "speedtest_raw" ];
65
+
source = ''
66
+
r = parse_json!(.message)
67
+
s = object!(r.server)
68
+
. = {
69
+
"download": to_float!(s.dl_speed) * 8,
70
+
"upload": to_float!(s.ul_speed) * 8,
71
+
"latency": to_float!(s.latency) / 1000000000,
72
+
"jitter": to_float!(s.jitter) / 1000000000,
73
+
"server": string!(s.sponsor) + " " + string!(s.name),
74
+
"server_id": string!(s.id),
75
+
}
76
+
'';
77
+
drop_on_error = true;
78
+
};
79
+
transforms.speedtest_metrics = {
80
+
type = "log_to_metric";
81
+
inputs = [ "speedtest" ];
82
+
metrics =
83
+
map
84
+
(m: {
85
+
type = "gauge";
86
+
field = m.field;
87
+
name = "speedtest_${m.name}";
88
+
tags = {
89
+
host = config.networking.hostName;
90
+
server = "{{server}}";
91
+
server_id = "{{server_id}}";
92
+
};
93
+
})
94
+
[
95
+
{
96
+
field = "download";
97
+
name = "download_bits_per_second";
98
+
}
99
+
{
100
+
field = "upload";
101
+
name = "upload_bits_per_second";
102
+
}
103
+
{
104
+
field = "latency";
105
+
name = "latency_seconds";
106
+
}
107
+
{
108
+
field = "jitter";
109
+
name = "jitter_seconds";
110
+
}
111
+
];
112
+
};
113
+
};
114
+
}
+113
configs/gg23/internet.nix
+113
configs/gg23/internet.nix
···
1
+
# Why is the gg23 internet slow? Runs on acheron, the gg23 router, whose WAN
2
+
# (et0) sits behind the Vodafone Station (ARRIS TG3442DE cable modem, LAN
3
+
# 192.168.0.1). Exports, via the Vector agent (configs/monitoring/agent.nix)
4
+
# into VictoriaMetrics on the monitoring servers:
5
+
# - the Station's DOCSIS channel power/SNR/lock and its status, scraped from
6
+
# its web UI (tools/vodafone-station-exporter)
7
+
# - latency and loss to the Station, the CMTS and the internet
8
+
# (smokeping_prober)
9
+
# - hourly speedtests (./internet-speedtest.nix)
10
+
# WAN throughput is in the host metrics already (et0 on acheron). Grafana
11
+
# shows it as "Internet", Home Assistant as the "Internet" dashboard
12
+
# (./internet-hass.nix).
13
+
{
14
+
config,
15
+
lib,
16
+
pkgs,
17
+
self,
18
+
...
19
+
}:
20
+
let
21
+
exporterPort = 9764;
22
+
smokeping = config.services.prometheus.exporters.smokeping;
23
+
24
+
# target label -> address. `cmts` is the first hop behind the cable modem
25
+
# (91.64.96.254 = gateway of the WAN 91.64.96.0/24, from mtr). If Vodafone
26
+
# moves us to another subnet it shows as 100% loss here: update it then.
27
+
targets = {
28
+
station = "192.168.0.1";
29
+
cmts = "91.64.96.254";
30
+
cloudflare = "1.1.1.1";
31
+
google = "8.8.8.8";
32
+
cloudflare6 = "2606:4700:4700::1111";
33
+
google6 = "2001:4860:4860::8888";
34
+
};
35
+
in
36
+
{
37
+
imports = [
38
+
./internet-hass.nix
39
+
./internet-speedtest.nix
40
+
];
41
+
42
+
clan.core.vars.generators.vodafone-station.prompts.password = {
43
+
description = "Vodafone Station (192.168.0.1) web UI admin password, passage admin/gg23/vodafone-router/pass";
44
+
type = "hidden";
45
+
persist = true;
46
+
};
47
+
48
+
systemd.services.vodafone-station-exporter = {
49
+
description = "Prometheus exporter for the Vodafone Station";
50
+
wantedBy = [ "multi-user.target" ];
51
+
after = [ "network-online.target" ];
52
+
wants = [ "network-online.target" ];
53
+
serviceConfig = {
54
+
ExecStart = lib.escapeShellArgs [
55
+
(lib.getExe self.packages.${pkgs.stdenv.hostPlatform.system}.vodafone-station-exporter)
56
+
"--listen"
57
+
"127.0.0.1:${toString exporterPort}"
58
+
"--password-file"
59
+
"%d/password"
60
+
];
61
+
LoadCredential = [
62
+
"password:${config.clan.core.vars.generators.vodafone-station.files.password.path}"
63
+
];
64
+
DynamicUser = true;
65
+
Restart = "on-failure";
66
+
RestartSec = 10;
67
+
};
68
+
};
69
+
70
+
services.prometheus.exporters.smokeping = {
71
+
enable = true;
72
+
listenAddress = "127.0.0.1";
73
+
hosts = lib.attrValues targets;
74
+
};
75
+
76
+
monitoring.extraMetrics = [ "internet_metrics" ];
77
+
78
+
services.vector.settings = {
79
+
# The Station is slow: one login + three pages take ~6s. Every scrape
80
+
# logs in and out again, so keep the interval relaxed.
81
+
sources.vodafone_station_raw = {
82
+
type = "prometheus_scrape";
83
+
endpoints = [ "http://127.0.0.1:${toString exporterPort}/metrics" ];
84
+
scrape_interval_secs = 60;
85
+
scrape_timeout_secs = 45;
86
+
};
87
+
sources.smokeping_raw = {
88
+
type = "prometheus_scrape";
89
+
endpoints = [
90
+
"http://127.0.0.1:${toString smokeping.port}${smokeping.telemetryPath}"
91
+
];
92
+
scrape_interval_secs = 30;
93
+
};
94
+
# smokeping_prober labels the probed address `host`, which is our
95
+
# machine label everywhere else: move it to `target` (plus the name from
96
+
# `targets`), then set host like the agent does.
97
+
transforms.internet_metrics = {
98
+
type = "remap";
99
+
inputs = [
100
+
"vodafone_station_raw"
101
+
"smokeping_raw"
102
+
];
103
+
source = ''
104
+
names = ${builtins.toJSON (lib.mapAttrs' (n: v: lib.nameValuePair v n) targets)}
105
+
if exists(.tags.host) {
106
+
.tags.target = del(.tags.host)
107
+
.tags.target_name = get(names, [.tags.target]) ?? .tags.target
108
+
}
109
+
.tags.host = "${config.networking.hostName}"
110
+
'';
111
+
};
112
+
};
113
+
}
+1
machines/acheron/config.nix
+1
machines/acheron/config.nix
+308
configs/monitoring/dashboards/internet.json
+308
configs/monitoring/dashboards/internet.json
···
1
+
{
2
+
"uid": "internet",
3
+
"title": "Internet (gg23)",
4
+
"schemaVersion": 39,
5
+
"time": {
6
+
"from": "now-24h",
7
+
"to": "now"
8
+
},
9
+
"refresh": "1m",
10
+
"panels": [
11
+
{
12
+
"id": 1,
13
+
"type": "timeseries",
14
+
"title": "Latency (avg, 1m)",
15
+
"gridPos": {
16
+
"x": 0,
17
+
"y": 0,
18
+
"w": 12,
19
+
"h": 9
20
+
},
21
+
"datasource": {
22
+
"type": "victoriametrics-metrics-datasource",
23
+
"uid": "victoriametrics"
24
+
},
25
+
"fieldConfig": {
26
+
"defaults": {
27
+
"unit": "s"
28
+
}
29
+
},
30
+
"targets": [
31
+
{
32
+
"refId": "A",
33
+
"expr": "rate(smokeping_response_duration_seconds_sum{host=\"acheron\"}[1m]) / rate(smokeping_response_duration_seconds_count{host=\"acheron\"}[1m])",
34
+
"legendFormat": "{{target_name}}"
35
+
}
36
+
]
37
+
},
38
+
{
39
+
"id": 2,
40
+
"type": "timeseries",
41
+
"title": "Latency p95 (5m)",
42
+
"gridPos": {
43
+
"x": 12,
44
+
"y": 0,
45
+
"w": 12,
46
+
"h": 9
47
+
},
48
+
"datasource": {
49
+
"type": "victoriametrics-metrics-datasource",
50
+
"uid": "victoriametrics"
51
+
},
52
+
"fieldConfig": {
53
+
"defaults": {
54
+
"unit": "s"
55
+
}
56
+
},
57
+
"targets": [
58
+
{
59
+
"refId": "A",
60
+
"expr": "histogram_quantile(0.95, sum(rate(smokeping_response_duration_seconds_bucket{host=\"acheron\"}[5m])) by (le, target_name))",
61
+
"legendFormat": "{{target_name}}"
62
+
}
63
+
]
64
+
},
65
+
{
66
+
"id": 3,
67
+
"type": "timeseries",
68
+
"title": "Packet loss (1m)",
69
+
"gridPos": {
70
+
"x": 0,
71
+
"y": 9,
72
+
"w": 12,
73
+
"h": 8
74
+
},
75
+
"datasource": {
76
+
"type": "victoriametrics-metrics-datasource",
77
+
"uid": "victoriametrics"
78
+
},
79
+
"fieldConfig": {
80
+
"defaults": {
81
+
"unit": "percentunit",
82
+
"min": 0
83
+
}
84
+
},
85
+
"targets": [
86
+
{
87
+
"refId": "A",
88
+
"expr": "1 - sum(rate(smokeping_response_duration_seconds_count{host=\"acheron\"}[1m])) by (target_name) / sum(rate(smokeping_requests_total{host=\"acheron\"}[1m])) by (target_name)",
89
+
"legendFormat": "{{target_name}}"
90
+
}
91
+
]
92
+
},
93
+
{
94
+
"id": 4,
95
+
"type": "timeseries",
96
+
"title": "WAN throughput (et0)",
97
+
"gridPos": {
98
+
"x": 12,
99
+
"y": 9,
100
+
"w": 12,
101
+
"h": 8
102
+
},
103
+
"datasource": {
104
+
"type": "victoriametrics-metrics-datasource",
105
+
"uid": "victoriametrics"
106
+
},
107
+
"fieldConfig": {
108
+
"defaults": {
109
+
"unit": "bps"
110
+
}
111
+
},
112
+
"targets": [
113
+
{
114
+
"refId": "A",
115
+
"expr": "rate(host_network_receive_bytes_total{host=\"acheron\",device=\"et0\"}[1m]) * 8",
116
+
"legendFormat": "down"
117
+
},
118
+
{
119
+
"refId": "B",
120
+
"expr": "-rate(host_network_transmit_bytes_total{host=\"acheron\",device=\"et0\"}[1m]) * 8",
121
+
"legendFormat": "up"
122
+
}
123
+
]
124
+
},
125
+
{
126
+
"id": 5,
127
+
"type": "timeseries",
128
+
"title": "Downstream power",
129
+
"gridPos": {
130
+
"x": 0,
131
+
"y": 17,
132
+
"w": 8,
133
+
"h": 9
134
+
},
135
+
"datasource": {
136
+
"type": "victoriametrics-metrics-datasource",
137
+
"uid": "victoriametrics"
138
+
},
139
+
"fieldConfig": {
140
+
"defaults": {
141
+
"unit": "suffix: dBmV"
142
+
}
143
+
},
144
+
"targets": [
145
+
{
146
+
"refId": "A",
147
+
"expr": "vodafone_station_docsis_power_dbmv{host=\"acheron\",direction=\"downstream\"}",
148
+
"legendFormat": "{{channel_type}} {{frequency_mhz}} MHz"
149
+
}
150
+
]
151
+
},
152
+
{
153
+
"id": 6,
154
+
"type": "timeseries",
155
+
"title": "Downstream SNR",
156
+
"gridPos": {
157
+
"x": 8,
158
+
"y": 17,
159
+
"w": 8,
160
+
"h": 9
161
+
},
162
+
"datasource": {
163
+
"type": "victoriametrics-metrics-datasource",
164
+
"uid": "victoriametrics"
165
+
},
166
+
"fieldConfig": {
167
+
"defaults": {
168
+
"unit": "suffix: dB"
169
+
}
170
+
},
171
+
"targets": [
172
+
{
173
+
"refId": "A",
174
+
"expr": "vodafone_station_docsis_snr_db{host=\"acheron\",direction=\"downstream\"}",
175
+
"legendFormat": "{{channel_type}} {{frequency_mhz}} MHz"
176
+
}
177
+
]
178
+
},
179
+
{
180
+
"id": 7,
181
+
"type": "timeseries",
182
+
"title": "Upstream power",
183
+
"gridPos": {
184
+
"x": 16,
185
+
"y": 17,
186
+
"w": 8,
187
+
"h": 9
188
+
},
189
+
"datasource": {
190
+
"type": "victoriametrics-metrics-datasource",
191
+
"uid": "victoriametrics"
192
+
},
193
+
"fieldConfig": {
194
+
"defaults": {
195
+
"unit": "suffix: dBmV"
196
+
}
197
+
},
198
+
"targets": [
199
+
{
200
+
"refId": "A",
201
+
"expr": "vodafone_station_docsis_power_dbmv{host=\"acheron\",direction=\"upstream\"}",
202
+
"legendFormat": "{{channel_type}} {{frequency_mhz}} MHz"
203
+
}
204
+
]
205
+
},
206
+
{
207
+
"id": 8,
208
+
"type": "timeseries",
209
+
"title": "Modulation",
210
+
"gridPos": {
211
+
"x": 0,
212
+
"y": 26,
213
+
"w": 12,
214
+
"h": 8
215
+
},
216
+
"datasource": {
217
+
"type": "victoriametrics-metrics-datasource",
218
+
"uid": "victoriametrics"
219
+
},
220
+
"fieldConfig": {
221
+
"defaults": {}
222
+
},
223
+
"targets": [
224
+
{
225
+
"refId": "A",
226
+
"expr": "vodafone_station_docsis_modulation_qam{host=\"acheron\"}",
227
+
"legendFormat": "{{direction}} {{channel_type}} {{frequency_mhz}} MHz"
228
+
}
229
+
]
230
+
},
231
+
{
232
+
"id": 9,
233
+
"type": "timeseries",
234
+
"title": "Station",
235
+
"gridPos": {
236
+
"x": 12,
237
+
"y": 26,
238
+
"w": 12,
239
+
"h": 8
240
+
},
241
+
"datasource": {
242
+
"type": "victoriametrics-metrics-datasource",
243
+
"uid": "victoriametrics"
244
+
},
245
+
"fieldConfig": {
246
+
"defaults": {}
247
+
},
248
+
"targets": [
249
+
{
250
+
"refId": "A",
251
+
"expr": "vodafone_station_up{host=\"acheron\"}",
252
+
"legendFormat": "scrape ok"
253
+
},
254
+
{
255
+
"refId": "B",
256
+
"expr": "vodafone_station_docsis_online{host=\"acheron\"}",
257
+
"legendFormat": "DOCSIS online"
258
+
},
259
+
{
260
+
"refId": "C",
261
+
"expr": "count(vodafone_station_docsis_locked{host=\"acheron\"} == 0) or vector(0)",
262
+
"legendFormat": "unlocked channels"
263
+
},
264
+
{
265
+
"refId": "D",
266
+
"expr": "vodafone_station_uptime_seconds{host=\"acheron\"} / 86400",
267
+
"legendFormat": "uptime (days)"
268
+
}
269
+
]
270
+
},
271
+
{
272
+
"id": 10,
273
+
"type": "timeseries",
274
+
"title": "Speedtest (hourly)",
275
+
"gridPos": {
276
+
"x": 0,
277
+
"y": 34,
278
+
"w": 24,
279
+
"h": 8
280
+
},
281
+
"datasource": {
282
+
"type": "victoriametrics-metrics-datasource",
283
+
"uid": "victoriametrics"
284
+
},
285
+
"fieldConfig": {
286
+
"defaults": {
287
+
"unit": "bps",
288
+
"custom": {
289
+
"drawStyle": "points",
290
+
"pointSize": 6
291
+
}
292
+
}
293
+
},
294
+
"targets": [
295
+
{
296
+
"refId": "A",
297
+
"expr": "speedtest_download_bits_per_second{host=\"acheron\"}",
298
+
"legendFormat": "down {{server}}"
299
+
},
300
+
{
301
+
"refId": "B",
302
+
"expr": "speedtest_upload_bits_per_second{host=\"acheron\"}",
303
+
"legendFormat": "up {{server}}"
304
+
}
305
+
]
306
+
}
307
+
]
308
+
}