Deploy SmokePing on a VPS

SmokePing does periodic latency/loss measurement (ping, or other probes) and renders RRD-backed graphs. Target: Ubuntu 24.04 LTS VPS, 1 vCPU / 1GB RAM is fine for dozens of targets at moderate polling intervals.

1. Base packages

shell

apt update && apt -y upgrade
apt -y install smokeping nginx fcgiwrap speedy-cgi-perl librrds-perl

The Ubuntu package installs to /etc/smokeping/, RRDs to /var/lib/smokeping/, and sets up an smokeping system user automatically. It also ships with /usr/sbin/smokeping as a daemon and a default systemd unit.

2. General config

/etc/smokeping/config.d/General:

shell

*** General ***

owner    = RamNode NOC
contact  = noc@ramnode.com
mailhost = localhost
sendmail = /usr/sbin/sendmail
imgcache = /var/lib/smokeping/cache
imgurl   = ../smokeping-cgi/cache
datadir  = /var/lib/smokeping/data
piddir   = /var/run/smokeping
smokemail = /etc/smokeping/smokemail.dist
cgiurl   = http://smokeping.example.com/smokeping-cgi/smokeping.cgi

# CDF percentile slope
slaves =

3. Presentation (thresholds, graph look)

/etc/smokeping/config.d/Presentation:

shell

*** Presentation ***

template = /etc/smokeping/basepage.html
charset  = utf-8

+ charts
menu = Charts
title = The most interesting graphs

++ loss
sorter = Loss(dsc)
menu = Top Loss
title = Top Loss

++ ping
sorter = Median(dsc)
menu = Top Ping
title = Top Ping

+ overview
width = 600
height = 50
range = 10h

++ mymenu
menu = Round Trip Times
title = Round Trip Time Statistics

+ detail
width = 800
height = 200
unison_tolerance = 2

"Last 3 Hours"    3h
"Last 30 Hours"   30h
"Last 10 Days"    10d
"Last 400 Days"   400d

4. Probes

/etc/smokeping/config.d/Probes:

shell

*** Probes ***

+ FPing
binary = /usr/bin/fping
step  = 300
offset = 50%
pings = 20

+ DNS
binary = /usr/bin/dig
pings  = 5
step   = 300

+ TCPPing
binary = /usr/bin/tcpping
pings  = 5
step   = 300

Install fping if not already present:

shell

apt -y install fping

5. Targets

/etc/smokeping/config.d/Targets:

shell

*** Targets ***

probe = FPing

menu = Top
title = Network Latency Overview
remark = RamNode monitored targets

+ Datacenters

menu = Datacenters
title = Datacenter Uplinks

++ dc_atl
menu = Atlanta uplink
title = Atlanta DC Uplink
host = gw-atl.ramnode.internal

++ dc_ny
menu = New York uplink
title = New York DC Uplink
host = gw-ny.ramnode.internal

+ Customers

menu = Customer Nodes
title = Customer-facing Nodes

++ kvm01
menu = kvm01
title = KVM Hypervisor 01
host = kvm01.ramnode.com

++ nlxsvz71
menu = nlxsvz71
title = OpenVZ dest node
host = nlxsvz71.ramnode.com

Add as many ++ target stanzas as you have hosts. Nested + Group sections build the menu tree.

6. Validate config

shell

smokeping --check --config=/etc/smokeping/config

7. Start the daemon

shell

systemctl enable --now smokeping
systemctl status smokeping

RRDs start populating in /var/lib/smokeping/data/ — give it a couple of polling cycles (5 min default step) before graphs show real data.

8. Web frontend (CGI via nginx + fcgiwrap)

The Debian/Ubuntu smokeping package installs the CGI to /usr/lib/smokeping/cgi-bin/smokeping.cgi (path may vary — confirm with dpkg -L smokeping | grep cgi).

Enable fcgiwrap:

shell

systemctl enable --now fcgiwrap

/etc/nginx/sites-available/smokeping:

shell

server {
    listen 80;
    server_name smokeping.example.com;

    location /smokeping-cgi/ {
        alias /usr/lib/smokeping/;  # confirm actual path with dpkg -L smokeping
        gzip off;

        location ~ \.cgi$ {
            include fastcgi_params;
            fastcgi_pass unix:/run/fcgiwrap.socket;
            fastcgi_param SCRIPT_FILENAME $request_filename;
        }
    }

    location / {
        return 302 /smokeping-cgi/smokeping.cgi;
    }
}

shell

ln -s /etc/nginx/sites-available/smokeping /etc/nginx/sites-enabled/
nginx -t && systemctl reload nginx
certbot --nginx -d smokeping.example.com

Add basic auth if this shouldn't be public:

shell

apt -y install apache2-utils
htpasswd -c /etc/nginx/.smokeping-htpasswd noc

shell

location /smokeping-cgi/ {
    auth_basic "SmokePing";
    auth_basic_user_file /etc/nginx/.smokeping-htpasswd;
    ...
}

9. Alerting (optional)

/etc/smokeping/config.d/Alerts:

shell

*** Alerts ***
to = noc@ramnode.com
from = smokeping@ramnode.com

+ loss20
type = loss
pattern = >0%,*12*
comment = loss on at least 1 of the last 12 probes

+ dupPL
type = loss
pattern = ==0%,==0%,>0%,>0%,>0%,>0%,>0%
comment = loss pattern following a change

Attach an alerts = loss20 line under any target's Targets stanza to apply.

10. Firewall

shell

ufw allow 80/tcp
ufw allow 443/tcp
ufw enable

11. Cross-check against Nagios

You already have Nagios 4 with check_public_activity and check_cpu.py ported. SmokePing is complementary — it shows latency/loss trends, Nagios gives pass/fail thresholds. Consider a Nagios passive check that reads the SmokePing RRD via rrdtool fetch if you want threshold alerts driven off the same latency data instead of duplicating fping checks in NRPE.

Post-install checklist

Confirm fping runs setuid or has CAP_NET_RAW (Ubuntu package usually handles this via dpkg-statoverride; check with getcap /usr/bin/fping)
Lock down /smokeping-cgi/ with basic auth or IP allowlist
Tune step/pings per probe if you want tighter granularity on critical uplinks
Set up RRD backups (/var/lib/smokeping/data/) — losing these loses historical trend data