Reducing GPSLogger request count

[hashworks]

GPSLogger currently has two ways to log locations to reitti with the custom URL method:

1. "Log to custom URL", which sends a request whenever a location point is captured.

This depends on the logging interval, where reitti recommends 30s. This method resulted in a high battery usage on my end since GPSLogger has to perform an HTTP Request every 30s.

2. "Allow auto sending", which sends stored locations in bulk

This depends on the configured auto-send-interval ("How often?"), which is 5 min on my end since I want the data in reitti to be somewhat up-to-date. This should save battery since the network stuff only needs to be woken every 5 minutes.

Using the defaults here has the opposite effect, since the way this works is that it takes the last CSV file and performs one HTTP request per line. The default filename includes the current day, so in the worst case this sends 86400s/30s=2880 requests, which stresses both the phone and the reitti server.

I was able to reduce the load by adding the current hour to the CSV filename, which results in 120 requests in the worst case, but it still sums up to 780 requests per hour.

Am I missing something here? Is it possible to configure GPSLogger to send the whole CSV to reitti or to avoid duplicate location log transfers at all?

[dgraf-gh]

I also noticed that behaviour but did gave that much about it. I will try to see what can be done in reitti but this will propably also need some information I have to look up about gps logger. It has the ability to also send a zip. This would reduce the requests to one but i need to find how it decides to do that instead of line by line

[hashworks]

It seems like the custom URL feature is only able to send line by line:

The 'Allow auto sending' toggle will also enable CSV logging out of necessity, and this feature will then make many requests to the Custom URL, one for each line in the recorded CSV. It does this at the auto-send interval which is by default 60 minutes. If you are writing these points to a database for example, it is your responsibility to 'deduplicate' when receiving these requests, for that reason it's a good idea to send a timestamp along.

[hashworks]

Maybe I'll spin up a new integration that provides an SFTP endpoint or watches a local folder. The SFTP auto send target can receive zips or individual gpx files:

$ sudo ls -al /var/lib/reitti/1
total 24
drwxrws--- 2 foo reitti  101 26. Okt 15:31 .
drwxr-xr-x 3 root          root     52 26. Okt 15:06 ..
-rw-r--r-- 1 foo reitti 4506 26. Okt 15:31 20251026T15.csv
-rw-r--r-- 1 foo reitti 5344 26. Okt 15:31 20251026T15.gpx
-rw-r--r-- 1 foo reitti 1683 26. Okt 15:26 20251026T15.zip
-rw-r--r-- 1 foo reitti   26 26. Okt 15:19 gpslogger_test.xml
$ sudo unzip -l /var/lib/reitti/1/20251026T15.zip
Archive:  /var/lib/reitti/1/20251026T15.zip
  Length      Date    Time    Name
---------  ---------- -----   ----
     3623  2025-10-26 15:26   20251026T15.gpx
     2948  2025-10-26 15:26   20251026T15.csv
---------                     -------
     6571                     2 files

This still sends duplicates, but it might be cheaper than line-by-line.

[dgraf-gh]

You could write a script to sends the gpx file to reitti. There is an endpoint for it.
https://www.dedicatedcode.com/projects/reitti/api/gpx-import/

If you send it multiple times during the day there is no problem, reitti will filter out already processed data.

So, let gpslogger send it to your server via sftp or owncloud or whatever works for you. Have a cron job or some other forms of automation which sends the gpx file to reitti.

I would not add sftp or some filewatcher to reitti. From my experience booth are to complex to add as a feature.

[hashworks]

Yeah, you are right. Simple enough.

#!/bin/bash

REITTI_WATCH_DIR="${REITTI_WATCH_DIR:-/var/lib/reitti/1}"

if [[ -z "$REITTI_ENDPOINT" ]]; then
  echo "Error: REITTI_ENDPOINT is not set."
  exit 1
fi

if [[ -z "$REITTI_TOKEN" ]]; then
  echo "Error: REITTI_TOKEN is not set."
  exit 1
fi

if [ ! -d "$REITTI_WATCH_DIR" ]; then
  echo "Error: Directory $REITTI_WATCH_DIR does not exist."
  exit 1
fi

inotifywait -m -e close_write "$REITTI_WATCH_DIR" --format '%w%f' |
while read FILE; do
  if [[ ! -f "$FILE" ]]; then
    continue
  fi

  case "$FILE" in
    *.zip)
      echo "Unzipping and removing $FILE"
      (cd "$REITTI_WATCH_DIR" && unzip -o "$FILE" >/dev/null && rm -f "$FILE")
      ;;
    *.gpx)
      echo "Importing $FILE"
      curl -s -X POST -H "X-API-TOKEN: ${REITTI_TOKEN}" -F "file=@${FILE}" "${REITTI_ENDPOINT}/api/v1/gpx/import" | jq -r '
        if .success == true then
          "Points scheduled: \(.pointsScheduled)"
        else
          if .message then
            "Import failed: \(.message)"
          else
            "Import failed due to unknown error."
          end
        end'
        ;;
    *)
      echo "Ignoring $FILE (unknown file type)"
      ;;
  esac
done

[dgraf-gh]

That looks really good.

Do you mind if I add this as an example to the Reitti Docs?

[hashworks]

Sure, go ahead. Depends on jq and inotify-tools.