Skip to content

Comcast Modem Power Level Monitor

Published: 2019-11-19

Below you'll find scripts to monitor a Comcast modem's power levels to help troubleshoot line problems.

The power of this tool becomes apparent when you use the data gathered to maintain an event journal of everything that happens related to your connection.

For example, for my business, everytime I show packet loss in my firewall I can corelate it to abnormal power levels in the modem. This is critical, not only to know that the problem lies outside of my control, but also to aide the Comcast technitians in fixing the issue. The graphs are especially usefull to show connection issues over time.

Windows Subsystem for Linux

Built-in linux system in Windows 10. You can run these scripts on any linux based system, but this was readily available and already on our existing network.

Start Cron & SSH at Windows Startup

Short batch file startup.bat used by a scheduled task in Windows to start the ssh and cron daemons in WSL. This is useful in case the host comptuer is restarted. (Windows Updates restarts mainly...)

1
2
3
4
5
6
@echo off

wsl -u root service ssh start
wsl -u root service cron start

exit

SSH

Having SSH installed in WSL makes it a lot easier to connect in and work on the script, check logs, etc.

Cron

  • Run the graph generating script every hour.
  • At 1 min intervals, check the python script hasn't errored out. If it's not running, start it.
1
2
0 * * * * /home/user/scripts/comcast/cm_run.sh
*/1 * * * * /home/user/scripts/comcast/keep_alive.sh

cm_rush.sh

Simple script to run the graph generating code. Could run R script directly from cron, but this allows other scripts to be called if needed.

1
2
3
#!/bin/bash

~/scripts/comcast/comcastplot.R

keep_alive.sh

Script to check if the python code is running. If it isn't, it will restart it.

1
2
3
4
5
6
#!/bin/bash

if ! pgrep -f 'cmpwrlevel2.py'
then
    nohup $(which python3) ~/scripts/comcast/cmpwrlevel2.py &
fi

Modem Power Levels

Poll the modem every minute and log the upstream (U) and downstream (D) powerlevels.

Acceptable ranges: (Note the DOCSIS 3.1, which is neede for higher speeds has lower tolerances.) https://www.dslreports.com/faq/16085

Sample Data

1
2
3
4
5
6
7
8
9
2019-09-16 01:30:17 U 43.78 dBmV D 3.97 dBmV
2019-09-16 01:31:18 U 43.78 dBmV D 3.97 dBmV
2019-09-16 01:32:18 U 43.78 dBmV D 3.97 dBmV
2019-09-16 01:33:18 U 43.78 dBmV D 3.96 dBmV
2019-09-16 01:34:19 U 43.78 dBmV D 3.96 dBmV
2019-09-16 01:35:18 U 43.78 dBmV D 4.03 dBmV
2019-09-16 01:36:18 U 43.78 dBmV D 4.00 dBmV
2019-09-16 01:37:18 U 43.78 dBmV D 3.98 dBmV
2019-09-16 01:38:17 U 43.78 dBmV D 3.97 dBmV

Get Data With Python

Script to get the data. The modem is quite "finicky".

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
#!/usr/bin/python3

# Comcast modem power levels reading script.
# Author: Eugenio C.
# Date: 09.16.2019

# TODO: Rewrite this script into a module with calls to objects.
# TODO: Put raw data in database. SQLite

# Load libraries
import logging
import re
import smtplib
import datetime
import pickle
import pathlib
import os
import sys
import time
import statistics
from email.message import EmailMessage
from bs4 import BeautifulSoup as bs
import requests

DATAFILE = 'cm_modem_pwr_levels.txt'
WAIT_TIME = 60

# ********************************************************************************
# Function Declerations
# ********************************************************************************

def send_text(msg_txt):
    msg = EmailMessage()

    msg.set_content(msg_txt)

    msg['Subject'] = 'CM Alarm'
    msg['From'] = 'python_script@yourdomain.com'
    msg['To'] = 'phonenumber@vtext.com'

    smtp = smtplib.SMTP('email.server.address')
    smtp.send_message(msg)
    smtp.quit()

# ********************************************************************************
# Main Script
# ********************************************************************************

ownp = os.path.abspath(os.path.dirname(sys.argv[0]))
tdnow = datetime.datetime.now().strftime("%Y%m%d")

# Setup Logging
logging.basicConfig(
    format='%(asctime)s:%(levelname)s:%(message)s',
    level=logging.DEBUG,
    filename=ownp + '/cmpwrlevel2.log')


# Get our "browser"
s = requests.Session()

# Reload cookies from previous session if they exist
if pathlib.Path(ownp + '/cookiefile2').exists():
    with open(ownp + '/cookiefile2', 'rb') as f:
        s.cookies.update(pickle.load(f))

# Set url and authentication
url = "http://10.1.10.1/check.php"
payload = {
    "username": "cusadmin", # Public, well known, customer-side, default credentials
    "password": "highspeed"
}

# Set our regex
regex = r'"Please Login First\!\"'
modemin = False

# Login loop, tries to connect to the modem.
while not modemin:
    try:
        # Login to set cookie
        logging.info("Getting: %s" % url)
        r = s.post(url, data=payload)

        logging.info("Status Code: %i" % r.status_code)
        r.raise_for_status()

        if not re.search(regex, str(r.content)):
            logging.info("Ok, we're in. Let's stay logged on and get data!")

            send_text("Logged on!")

            modemin = True
        else:
            logging.info("No Luck! Let's try again in 10 seconds.")
            time.sleep(10)

    except Exception as e:
        logging.error(e)

# Save cookies
with open(ownp + '/cookiefile2', 'wb') as f:
    pickle.dump(s.cookies, f)

# Now we can access the page we need
url = "http://10.1.10.1/comcast_network.php"

getdata = True

# Let's start the data loop
while getdata:
    try:
        logging.info("Getting: %s" % url)
        r = s.get(url)

        logging.info("Status Code: %i" % r.status_code)
        r.raise_for_status()

        # Get the date/time close to the time we read the values
        currentDT = datetime.datetime.now()

        # Get the html
        bs_content = bs(r.content, "lxml")

        # Narrow the html down to the two tables
        down_data = bs_content.find_all('table')[0].get_text()
        up_data = bs_content.find_all('table')[1].get_text()

        # Set our regex
        regex = r"[-0-9\.]* dBmV"

        # Get the power levels
        d_matches = re.finditer(regex, down_data, re.MULTILINE)
        u_matches = re.finditer(regex, up_data, re.MULTILINE)

        d_pwr_sum = []
        u_pwr_sum = []

        # For each power level remove the label and convert to a number
        # Sum up all the power levels so we can divide later for the average
        for d_matchNum, match in enumerate(d_matches, start=1):
            d_pwr = float(match.group().replace(" dBmV", ""))
            d_pwr_sum.append(d_pwr)

        for u_matchNum, match in enumerate(u_matches, start=1):
            u_pwr = float(match.group().replace(" dBmV", ""))
            u_pwr_sum.append(u_pwr)

        # Print info to log file
        logging.info("DW: {}".format(','.join(map(str, d_pwr_sum))))
        logging.info("DW: H%.1f, L%.1f, M%.1f, A%.2f" %
                     (max(d_pwr_sum),
                      min(d_pwr_sum),
                      statistics.median(d_pwr_sum),
                      statistics.mean(d_pwr_sum)))

        logging.info("UP: {}".format(','.join(map(str, u_pwr_sum))))
        logging.info("UP: H%.1f, L%.1f, M%.1f, A%.2f" %
                     (max(u_pwr_sum),
                      min(u_pwr_sum),
                      statistics.median(u_pwr_sum),
                      statistics.mean(u_pwr_sum)))

        # Calculate the averages
        avg_u_pwr = statistics.mean(u_pwr_sum)
        avg_d_pwr = statistics.mean(d_pwr_sum)

        # If the power levels are outside of spec, send an alert text message
        if (avg_u_pwr > 49) or (avg_u_pwr < 40) or (avg_d_pwr > 10) or (avg_d_pwr < -10):
            send_text("U%.2f D%.2f" % (avg_u_pwr, avg_d_pwr))

        # Write data to file
        logging.info("Writing data to: %s" % DATAFILE)
        with open(DATAFILE, 'a') as o:
            o.write(
                "%s U %.2f dBmV D %.2f dBmV\n" %
                (currentDT.strftime("%Y-%m-%d %H:%M:%S"),
                 avg_u_pwr,
                 avg_d_pwr))

        logging.info("Waiting %i seconds." % (WAIT_TIME))
        time.sleep(WAIT_TIME)

    except Exception as e:
        logging.error(e)

        send_text("Exception!")

        with open(ownp + '/content.html', 'w+') as h:
            h.write(r.text)

        # Exit loop, cron will restart
        getdata = False

Sample Graphs

Downstream Power Levels

Upstream Power Levels

Historical Graphs

This is where you can really show line issues over time. You can see why we had random drops at the begging of the graph. Some variance is expected on power levels, especially due to temperature, but no mor than 2-3 decibels.

You can see where Comast finally fixed the issue. You can also see where we added back our splitter for the TV.

Note: the gaps in the line is missing data. The modem sometimes does not repond to web requests. After some time, it will now allow web logins, yet somehow, with the python script and reusing cookies we are allowed in to collect data. Some times the page will refresh without the signal levels. Other times, during line problems the page will not be available.

Click here to see the full image...

Click here to see the full image...

Generate Graphs With R

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
#!/usr/bin/env Rscript

# Comcast modem power levels plotting script.
# Author: Eugenio Cilento
# Date: 09.16.2019

# Load libraries
library(ggplot2, warn.conflicts = FALSE)
library(scales, warn.conflicts = FALSE)
library(ggrepel, warn.conflicts = FALSE)
library(lubridate, warn.conflicts = FALSE)
library(dplyr, warn.conflicts = FALSE)

consts <- list(img_w = 14,
    img_h = 8.5,
    img_dpi = 320,
    img_q = 30,
    days = 6)

# ***************************************************************************************
# Get the data, clean it up and get it ready for use in the graph.
# ***************************************************************************************

# Read in the modem power levels from the text file.
pwr <- read.table("cm_modem_pwr_levels.txt")

pwr = pwr[c(-3,-5,-6,-8)] # Remove uneeded columns.

colnames(pwr) = c("Date","Time","UP","DW") # Let's give the columns some names.

pwr$DateTime <- paste(pwr$Date, pwr$Time) # Create a new timstamp column.

# Create a new column used for the facets. Basically to show the Days on the graph.
pwr$Dow <- strptime(pwr$DateTime, "%Y-%m-%d")
pwr$Dow <- format(pwr$Dow, "%m-%d %a")

pwr = pwr[c(-1,-2)] # Some more cleanup.

# Make that column an actual date time field
pwr$DateTime <- strptime(pwr$DateTime, "%Y-%m-%d %H:%M:%S")
pwr$DateTime <- as.POSIXct(pwr$DateTime, tz = "America/Chicago") # Convert to POSIXct type

pwr <- pwr[pwr$DateTime >= as.POSIXct(Sys.Date()-consts$days),] # Limit the data.

# When did the modem signal go outside of the acceptable range?
pwrovr <- pwr[which(pwr[,1] > 49 | pwr[,1] < 35 ),]

# Dataframe for ranges
dmax <- pwr %>% group_by(Dow) %>% summarise(Max = max(DateTime))
dmin <- pwr %>% group_by(Dow) %>% summarise(Min = min(DateTime))
dminmax <- merge(dmax,dmin,by = "Dow")

# Put in variables so we dont have to repeat code
dta = as.POSIXct(Sys.Date())
dtw = format(as.POSIXct(Sys.Date()), "%m-%d %a")

# ***************************************************************************************
# Beginning of the UPSTREAM graph script.
# ***************************************************************************************

# Dataframe for text labels for ranges
text_ann <- data.frame(UP = 49, DW = 5,
  lab = "Packet Loss\nBegins: +49", DateTime = dta, Dow = dtw)

text_ann <- rbind(text_ann, 
  data.frame(UP = 35, DW = 5, 
    lab = "Recommended:\n+ 35 - 47", DateTime = dta, Dow = dtw))

hour(text_ann$DateTime) = 2 #position of text within facet

# Create the plot
gg <- ggplot(pwr, aes(DateTime, UP)) +
  geom_point( # Points on a graph seem to work best for this type data.
    color = "black",
    size = .1
  ) + 
  facet_grid( # This gives the days in a nice column.
    col = vars(Dow),
    labeller = labeller(Dow = label_wrap_gen(5)), # Word rap the labels.
    space = 'free', 
    scales = 'free', 
    switch = 'x'
  ) +
  labs( # Needs some good titles and labels to make it easier to understand.
    title="Cable Modem Upstream Power", 
    subtitle="Comcast Business Internet", 
    y="Upstream Power Level (dBmV)", 
    x="Time (Date & Hour in 24Hr format)", 
    caption= paste( # Timestamp the graph.
      "International Assurance of TN, Inc. IT Department -",
      Sys.time()
    )
  ) +
  geom_rect( # color the acceptable range so it stands out.
    data = dminmax, 
    aes(
      xmin = Min, 
      xmax = Max, 
      ymin = 35, 
      ymax = 47), 
    alpha = 0.2, 
    fill="green", 
    inherit.aes = FALSE
  ) +
  geom_rect( # color the acceptable range so it stands out.
    data = dminmax, 
    aes(
      xmin = Min, 
      xmax = Max, 
      ymin = 47, 
      ymax = 49), 
    alpha = 0.2, 
    fill="red", 
    inherit.aes = FALSE
  ) +
  geom_text( # add labels for the ranges
    data = text_ann, 
    aes(x=DateTime, y=UP, label=lab), 
    vjust=-.2,
    hjust=0,
    size=2.5,
    color= "black"
  ) +
  geom_hline(yintercept=49, linetype="dashed", color = "red") +
  geom_hline(yintercept=47, linetype="dashed", color = "green") +
  geom_hline(yintercept=35, linetype="dashed", color = "green") +
  coord_cartesian( # Set the y scale. The range of modem signal levels.
    ylim = c(0, 60)
  ) +
  scale_y_continuous( # Set where the lines show on the graph.
    minor_breaks = seq(0 , 60, 1), 
    breaks = seq(0, 60, 10)
  ) +
  scale_x_datetime( # Set the x scale based on the hour of the day.
    breaks = pretty_breaks(), # Makes the hours fit nicely in a day.
    expand = c(0, 0), # Brings the panel/facets together, seamless transition from day to day.
    labels = date_format("%H", tz = Sys.timezone(location = TRUE)), 
    date_breaks = "2 hour",
    date_minor_breaks = "1 hour"
  ) + 
  theme_bw() + # Start off with a base theme, then modify it.
  theme(
    panel.border = element_rect(
      color = "dark gray"),
    panel.spacing.x = unit(0,"line"),
    strip.placement = "outside",
    strip.text.x = element_text(
      size = 6
    ),
    panel.grid.major = element_line(
      color = "light gray"),
    panel.grid.minor = element_line(
      color = "light gray"
    ),
    axis.text.x = element_text( # Hour of the day
      size = 6,
      angle = 90,
      vjust = 0.5,
      hjust = 0.5),
    axis.line = element_line( # Make the axis stand out.
      color = "darkblue", 
      size = 1, 
      linetype = "solid")
  ) 

if (dim(pwrovr)[1] != 0) { # Make sure we have out of range plots before adding them 
  gg = gg +
    geom_text_repel( # Label the out-of-range plots
      aes(label = paste(hour(DateTime),":", minute(DateTime),"-",UP)), 
      pwrovr,
      size = 2,
      color = "red"
    )
}

# Save the graph to a JPEG file. High resolution allows more clarity and printability.
ggsave("cm_modem_up_pwr_levels.jpg", 
       width = consts$img_w, 
       height = consts$img_h, 
       dpi = consts$img_dpi,
       quality = consts$img_q,
       limitsize = FALSE)

# ***************************************************************************************
# Beginning of the DOWNSTREAM graph script.
# ***************************************************************************************

# Dataframe for text labels for  ranges
text_ann2 <- data.frame(UP = -6, DW = 5,
  lab = "Recommended:\n-7 to 7", DateTime = dta, Dow = dtw)

text_ann2 <- rbind(text_ann2, 
  data.frame(UP = -9, DW = 5, 
    lab = "Acceptable:\n+- 7 to 10", DateTime = dta, Dow = dtw))

text_ann2 <- rbind(text_ann2, 
  data.frame(UP = -14, DW = 5, 
    lab = "Maximum:\n+- 10 to 15 ", DateTime = dta, Dow = dtw))

text_ann2 <- rbind(text_ann2, 
  data.frame(UP = -18, DW = 5, 
    lab = "Out of Spec:\n> +-15 ", DateTime = dta, Dow = dtw))

hour(text_ann2$DateTime) = 2 #position of text within facet

# Filter into a new table the out-of-range instances.
pwrovr <- pwr[which(pwr[,2] > 10 | pwr[,2] < -10 ),]

# Generate the downstream graph. Identical to upstream but with different ranges.
gg <- ggplot(pwr, aes(DateTime, DW)) +
  geom_point(
    color = "black",
    size = .1
  ) +
  facet_grid(
    col = vars(Dow),
    labeller = labeller(Dow = label_wrap_gen(5)),
    space = 'free', 
    scales = 'free', 
    switch = 'x'
  ) +
  coord_cartesian(
    ylim = c(-30, 20)
  ) +
  scale_y_continuous(
    minor_breaks = seq(-30 , 20, 1), 
    breaks = seq(-30, 20, 10)
  ) +
  labs(
    title="Cable Modem Downstream Power",
    subtitle="Comcast Business Internet",
    y="Downstream Power Level (dBmV)",
    x="Time (Date & Hour in 24Hr format)",
    caption= paste(
      "International Assurance of TN, Inc. IT Department -",
      Sys.time()
    )
  ) +
  geom_text( # write the labels for the ranges
    data = text_ann2, 
    aes(x=DateTime, y=UP, label=lab), 
    vjust=.3,
    hjust=0,
    size=2.5,
    color= "black"
  ) +
  geom_hline(yintercept=-15, linetype="dashed", color = "red") +
  geom_hline(yintercept=-10, linetype="dashed", color = "yellow") +
  geom_hline(yintercept=-7, linetype="dashed", color = "green") +
  geom_hline(yintercept=7, linetype="dashed", color = "green") +
  geom_hline(yintercept=10, linetype="dashed", color = "yellow") +
  geom_hline(yintercept=15, linetype="dashed", color = "red") +
  geom_rect( # color the acceptable range so it stands out.
    data = dminmax, 
    aes(
      xmin = Min, 
      xmax = Max, 
      ymin = -7, 
      ymax = 7), 
    alpha = 0.2, 
    fill="green", 
    inherit.aes = FALSE
  ) +
  geom_rect( # color the acceptable range so it stands out.
    data = dminmax, 
    aes(
      xmin = Min, 
      xmax = Max, 
      ymin = -10, 
      ymax = -7), 
    alpha = 0.2, 
    fill="yellow", 
    inherit.aes = FALSE
  ) +
  geom_rect( # color the acceptable range so it stands out.
    data = dminmax, 
    aes(
      xmin = Min, 
      xmax = Max, 
      ymin = 7, 
      ymax = 10), 
    alpha = 0.2, 
    fill="yellow", 
    inherit.aes = FALSE
  ) +
   geom_rect( # color the acceptable range so it stands out.
    data = dminmax, 
    aes(
      xmin = Min, 
      xmax = Max, 
      ymin = -15, 
      ymax = -10), 
    alpha = 0.2, 
    fill="red", 
    inherit.aes = FALSE
  ) +
  geom_rect( # color the acceptable range so it stands out.
    data = dminmax, 
    aes(
      xmin = Min, 
      xmax = Max, 
      ymin = 10, 
      ymax = 15), 
    alpha = 0.2, 
    fill="red", 
    inherit.aes = FALSE
  ) +
  scale_x_datetime(
    breaks = pretty_breaks(),
    expand = c(0, 0),
    labels = date_format("%H", tz = Sys.timezone(location = TRUE)),
    date_breaks = "2 hour",
    date_minor_breaks = "1 hour"
  ) +
  theme_bw() +
  theme(
    panel.border = element_rect(
      colour = "dark gray"),
    panel.spacing.x = unit(0,"line"),
    strip.placement = "outside",
    strip.text.x = element_text(
      size = 6
    ),
    panel.grid.major = element_line(
      colour = "light gray"),
    panel.grid.minor = element_line(
      color = "light gray"
    ),
    axis.text.x = element_text(
      size = 6,
      angle = 90,
      vjust = 0.5,
      hjust = 0.5),
    axis.line = element_line(
      colour = "darkblue", 
      size = 1, 
      linetype = "solid")
  )

if (dim(pwrovr)[1] != 0) {  # Make sure we have out of range before plotting
  gg = gg +
    geom_text_repel( # Label the out-of-range plots
      aes(label = paste(DateTime, "-", DW, "dBmV")),
      pwrovr,
      size = 2,
      color = "red"
    )
}

# Save graph to file.
ggsave("cm_modem_dwn_pwr_levels.jpg", 
       width = consts$img_w, 
       height = consts$img_h, 
       dpi = consts$img_dpi,
       quality = consts$img_q,
       limitsize = FALSE)