logging.md

Logging

The logging sub-system provides a way to peri.odically sample variables from the Crazyflie.

For more informaiton on how to use the logging and parameter sub-systems see this link

List available variables

cfcli log list

This will produce an output similar to this:

Name                           | Type 
-------------------------------|------
DTR_P2P.rx_state               | U8
DTR_P2P.tx_state               | U8
acc.x                          | F32
acc.y                          | F32
acc.z                          | F32
activeMarker.btSns             | U8
activeMarker.i2cOk             | U8
baro.asl                       | F32
...

Where the Name column is the variable name (group.name) and the Type column describes the data type of the variable.

For machine-readable output, add the global --csv flag:

cfcli --csv log list

The first line is a header row, then one variable per line:

name,type
DTR_P2P.rx_state,U8
DTR_P2P.tx_state,U8
acc.x,F32

Log variables

There's two ways to log variables, either specify exactly which variables to log on the command line, or leave it blank to get an interactive prompt with all the available variables where you can select which ones to log.

To print the values of the acc.x and acc.y variables at each 10 ms use the following command:

cfcli log print acc.x,acc.y -p 10

The period defaults to 100 ms if -p/--period is omitted.

This will produce an output similar to this:

LogData { timestamp: 377433600, data: {"acc.x": F32(0.02040638), "acc.y": F32(-0.011233097)} }
LogData { timestamp: 377436160, data: {"acc.x": F32(0.018765358), "acc.y": F32(-0.01344875)} }

CSV output

For piping into another tool or capturing to a file, use the global --csv flag. The first line is a header row (timestamp_ms plus one column per requested variable), and each sample is one row with plain numeric values:

cfcli --csv log print acc.x,acc.y -p 10

timestamp_ms,acc.x,acc.y
377433600,0.02040638,-0.011233097
377436160,0.018765358,-0.01344875

Rows are flushed per sample, so consumers see data in real time rather than in stdio-buffered chunks.

Stop streaming after a fixed duration

log print is a streaming command — by default it runs until the link is broken. Combine it with the global --timeout flag to stop cleanly after a fixed wall-clock duration:

cfcli --timeout 3000 log print acc.x,acc.y -p 10

When --timeout fires on a streaming command, the process exits 0 (the timer is the intended way to stop it). This is the recommended pattern when running cfcli log print from a script or CI step.

Set up base station (i.e set channel and download calib data?)