How to use
Add in your server :
# server
track_usage(storage_mode = store_json(path = "logs/"))
It will store logs in a sub-folder logs/
of your
application.
If needed (in complex application), you can force dependencies in UI with :
# UI
use_tracking()
Storage modes
Four modes are available:
-
store_json()
: store logs has separate JSON files (one by session). -
store_rds()
: store logs has individually RDS files (one by session). -
store_sqlite()
: store logs in a SQLite database. -
store_null()
: don’t write logs on disk, print them in R console. -
store_custom()
: use a custom function to save logs wherever you want. -
store_googledrive()
: store logs(as JSON files) in Googledrive.
Deployment method
Shiny server, RStudio Connect
On a server, if you want to save logs on disk, don’t forget to set write permission on the folder you want to save logs.
On RStudio Connect, you need to use an absolute path to specify the directory where to save logs. You can find more information here: Persistent Storage on RStudio Connect
ShinyProxy
With ShinyProxy, you can use
a Docker volume to write logs outside of the application container. In
application.yml
, you can something like this in the specs
describing the application:
container-volumes: [ "/var/log/shinylogs:/root/logs" ]
/var/log/shinylogs
is a directory on the server where
you deploy your applications with ShinyProxy. /root/logs
is
a directory inside your Docker image, it’s the path you can use in
track_usage()
, e.g. :
track_usage(
storage_mode = store_json(path = "/root/logs")
)
shinyapps.io
There’s no persistent data storage on shinyapps.io, so you can’t save
logs as JSON or RDS files, you have to use a remote storage method. You
can use the builtin method provided to send logs in Googledrive with
store_googledrive
or use store_custom
to send
logs wherever you want (dropbox, remote database, …).
To use Googledrive, you’ll need to work with Google’s API and set auth to your account, see {gargle} documentation for examples and how to.
Recorded informations
Session
Those data are metadata about the application and the user’s browser, here are the filed recorded :
- app : name of the application
- user : name of the user (if using Shiny-server pro for example)
- server_connected : timestamp of when application has been launched (server time)
- sessionid : a session ID to match the session with other recorded informations (inputs, outputs, errors)
- server_disconnected : timestamp of when the application was disconnected (server time)
- user_agent : browser user-agent
- screen_res : resolution of the user screen (width x height)
- browser_res : resolution of the user browser (width x height)
- pixel_ratio : pixel ratio of the browser
- browser_connected : timestamp of when application has been launched (browser time, depends on user timezone)
Example:
#> app user server_connected sessionid
#> 1 iris-cluster dreamRs 2019-06-19 14:07:09 9815194cfaa6317fbea68ae9537d63d1
#> 2 iris-cluster dreamRs 2019-06-19 14:56:32 0feeacf3201f5cca088059cec2b0a710
#> 3 iris-cluster dreamRs 2019-06-19 14:57:45 8b12c138f159cc5805e136338dddac59
#> 4 iris-cluster dreamRs 2019-06-19 15:03:00 f3950a1588b78d45c53c756a4136df67
#> 5 iris-cluster dreamRs 2019-06-19 15:08:53 254cafb3d5866384f13022d066546cae
#> 6 iris-cluster dreamRs 2019-08-04 10:58:47 8a431f4b90b3b1926047dd2539be0793
#> server_disconnected
#> 1 2019-06-19 14:07:17
#> 2 2019-06-19 14:57:39
#> 3 2019-06-19 15:01:36
#> 4 2019-06-19 15:08:13
#> 5 2019-06-19 15:09:38
#> 6 2019-08-04 10:58:49
#> user_agent
#> 1 Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/74.0.3729.169 Safari/537.36
#> 2 Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/74.0.3729.169 Safari/537.36
#> 3 Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/74.0.3729.169 Safari/537.36
#> 4 Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/74.0.3729.169 Safari/537.36
#> 5 Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/74.0.3729.169 Safari/537.36
#> 6 <NA>
#> screen_res browser_res pixel_ratio browser_connected
#> 1 1920x1080 1574x724 1 2019-06-19 14:07:09
#> 2 1920x1080 1920x937 1 2019-06-19 14:56:32
#> 3 1920x1080 1920x937 1 2019-06-19 14:57:45
#> 4 1920x1080 1920x937 1 2019-06-19 15:03:00
#> 5 1920x1080 1920x937 1 2019-06-19 15:08:53
#> 6 <NA> <NA> NA <NA>
Inputs
Data about inputs, by default all inputs are recorded (even those not define by developper, like with {htmlwidgets} : {DT}, {leaflet}, …)
- sessionid : the same ID as in session object
- name : inputId of the input
- timestamp : timestamp when the input has changed
- value : the value taken by the input (can be a list in case of complex input)
- type : type of input (if defined)
- binding : binding for the input (if defined)
Example:
#> sessionid name timestamp value
#> 1 9815194cfaa6317fbea68ae9537d63d1 xcol 2019-06-19 14:07:11 Sepal.Width
#> 2 9815194cfaa6317fbea68ae9537d63d1 clusters 2019-06-19 14:07:15 6
#> 3 9815194cfaa6317fbea68ae9537d63d1 clusters 2019-06-19 14:07:15 4
#> 4 9815194cfaa6317fbea68ae9537d63d1 clusters 2019-06-19 14:07:15 5
#> 5 9815194cfaa6317fbea68ae9537d63d1 clusters 2019-06-19 14:07:14 7
#> 6 9815194cfaa6317fbea68ae9537d63d1 clusters 2019-06-19 14:07:13 4
#> type binding
#> 1 <NA> shiny.selectInput
#> 2 shiny.number shiny.numberInput
#> 3 shiny.number shiny.numberInput
#> 4 shiny.number shiny.numberInput
#> 5 shiny.number shiny.numberInput
#> 6 shiny.number shiny.numberInput
Outputs
Data recorded each time an output is refreshed :
- sessionid : the same ID as in session object
- name : outputId of the output
- timestamp : timestamp when the output has been updated
- type : type of output (if defined)
- binding : binding for the output (if defined)
Example:
#> sessionid name timestamp binding
#> 1 9815194cfaa6317fbea68ae9537d63d1 plot1 2019-06-19 14:07:15 shiny.imageOutput
#> 2 9815194cfaa6317fbea68ae9537d63d1 plot1 2019-06-19 14:07:15 shiny.imageOutput
#> 3 9815194cfaa6317fbea68ae9537d63d1 plot1 2019-06-19 14:07:14 shiny.imageOutput
#> 4 9815194cfaa6317fbea68ae9537d63d1 plot1 2019-06-19 14:07:15 shiny.imageOutput
#> 5 9815194cfaa6317fbea68ae9537d63d1 plot1 2019-06-19 14:07:11 shiny.imageOutput
#> 6 9815194cfaa6317fbea68ae9537d63d1 plot1 2019-06-19 14:07:10 shiny.imageOutput
Errors
Errors are recorded only when propagated through an output, this is the red message users see in application, infos saved are:
- sessionid : the same ID as in session object
- name : outputId of the output where an error happened
- timestamp : timestamp of the error
- error : error message (if any)
- value : additional data for the error (generally NULL) (if defined)
Example:
#> sessionid name timestamp
#> 1 f8f50a3743023aae7d0d6350a2fd6841 plot1 2019-06-19 14:07:18
#> error
#> 1 NA/NaN/Inf in foreign function call (arg 1)