dAm2K Blog

Monitoring the health of your UPS (Uninterruptible Power Supply) is crucial to ensure your systems are protected from power outages and electrical anomalies. In this project, I built a Grafana dashboard that visualizes real-time UPS metrics collected via apcaccess and delivered to InfluxDB through Telegraf.

How It Works

The workflow consists of four key components:

1. apcaccess — The command-line tool that queries an APC UPS through the apcupsd daemon (via USB).
2. Custom Bash Script — Extracts and formats specific UPS metrics from apcaccess.
3. Telegraf — Executes the script, parses the output, and sends the data to InfluxDB.
4. Grafana — Reads from InfluxDB and displays the data in a visual, interactive dashboard.

The Bash Script

The script get_ups_values.sh calls apcaccess -u, which outputs UPS status in an easy-to-parse, colon-separated format.
Using awk, we clean and format these values into InfluxDB’s line protocol:

#!/bin/bash
KEYS=$(
apcaccess -u | awk -F: '
{
    gsub(/^[ \t]+|[ \t]+$/, "", $2)
}
/MODEL/      {out=out "model=" $2 ","}
/SERIALNO/   {out=out "serialno=" $2 ","}
/NOMPOWER/   {out=out "nompower=" $2 ","}
END {
    sub(/,$/, "", out)
    print out
}
'
)

VALS=$(
apcaccess -u | awk -F: '
{
    gsub(/^[ \t]+|[ \t]+$/, "", $2)
}
/STATUS/    {out=out "ups_status=\"" $2 "\","}
/LINEV/     {out=out "ups_input_volts=" $2 ","}
/OUTPUTV/   {out=out "ups_output_volts=" $2 ","}
/LOADPCT/   {out=out "ups_load_pct=" $2 ","}
/BCHARGE/   {out=out "ups_battery_charge=" $2 ","}
/TIMELEFT/  {out=out "ups_time_left_min=" $2 ","}
/TONBATT/   {out=out "ups_time_on_battery=" $2 ","}
/CUMONBATT/ {out=out "ups_total_time_on_battery=" $2 ","}
/NUMXFERS/  {out=out "ups_transfer_count=" $2 ","}
END {
    sub(/,$/, "", out)
    print out
}
'
)
echo "ups,host=$(uname -n),source=apcaccess,$KEYS $VALS"

Telegraf Configuration

In /etc/telegraf/telegraf.d/get_ups_values.conf, we add an exec input plugin to run the script at regular intervals:

[[inputs.exec]]
  commands = ["/usr/local/bin/get_ups_values.sh"]
  timeout = '30s'
  data_format = "influx"
  interval = "30s"

Every 10 seconds, Telegraf runs the script, captures the metrics, and sends them directly to InfluxDB.

The InfluxDB Data Model

The script’s output already matches InfluxDB’s line protocol format, making ingestion seamless.
Each metric is stored with tags and fields, making it easy to query in Grafana:

ups,host=firegate3,source=apcaccess,model=VP1600ELCD,nompower=960 ups_status="ONLINE",ups_input_volts=226.0,ups_load_pct=6.0,ups_battery_charge=100.0,ups_time_left_min=229.5,ups_output_volts=226.0,ups_transfer_count=4,ups_time_on_battery=0,ups_total_time_on_battery=480

The Grafana Dashboard

Once the data is in InfluxDB, Grafana can query it to display:

• UPS Status (ONLINE, ONBATT, etc.)
• Input & Output Voltage
• Load Percentage
• Battery Charge
• Time Left on Battery
• Number of Transfers to Battery

The result is a real-time dashboard that not only provides immediate visibility but also stores historical data for trend analysis.

Some of the benefits are:

• Proactive Monitoring — Detect issues before they become critical.
• Historical Insights — Spot patterns in power quality over time.
• Automation-Friendly — Integrates easily with alerts in Grafana.

Conclusion

With just a few lines of Bash, a Telegraf exec plugin, and a Grafana dashboard, you can transform raw UPS data into a rich monitoring experience. This setup is lightweight, flexible, and easy to expand with additional metrics or automation triggers.

You can get my grafana UPS dashboard json for free.

HTH.