posit-workbench-job-scheduler is a standalone R project that contains functions to assist with programmatically launching Workbench Jobs on a Kubernetes cluster, and scheduling these to run at a specific time and/or at regular intervals:
launch_workbench_job() launches a Workbench Job on the default Kubernetes clusterschedule_workbench_job() schedules a Workbench Job to launch on the default Kubernetes cluster at a specific time and/or at regular intervalscancel_workbench_job_schedule() cancels a schedule created by the function schedule_workbench_job()posit-workbench-job-scheduler is not an R package, and does not require installation. Instead, this repository should be cloned to a directory accessible from Posit Workbench.
posit-workbench-job-scheduler is a standalone R project that should be opened in a Posit Workbench session. This session needs to run continuously in the background for Workbench Jobs to be launched on a Kubernetes cluster at the right time.
It is recommended that posit-workbench-job-scheduler should be run in a session with 0.25 CPUs and 256 MiB of memory; this should comfortably handle 10+ individual schedules.
posit-workbench-job-scheduler will automatically start running when the project is opened in a new Posit Workbench session. Confirmation will be provided by the following message on the Console:
✔ Posit Workbench Job Scheduler is running
ℹ Please ensure that you do not quit this Posit Workbench session!The function launch_workbench_job() immediately launches a Workbench Job on the default Kubernetes cluster.
launch_workbench_job(
job_name,
project_path,
script,
n_cpu,
n_ram
)A descriptive name to assign to the Workbench Job. Default is the filename of script.
The full file path to the root directory of the project that script should be run in e.g. /path/to/project/.
A file path to the R script to be run by the Workbench Job, relative to project path e.g. code/script.R.
The number of CPUs to request for the Workbench Job e.g. 1.0. Default is the default number of CPUs for a session as defined in the Kubernetes cluster.
The amount of memory (RAM) in MiB to request for the Workbench Job e.g. 4096. Default is the default amount of memory for a session as defined in the Kubernetes cluster.
job_id <- launch_workbench_job(
job_name = "A name for this job",
project_path = here::here(),
script = "code/script.R",
n_cpu = 0.25,
n_ram = 128
)
✔ A Workbench Job with id 'S3ViZXJuZXRlczp0ZXN0LXdvcmtiZW5jaC1qb2ItZnhzOWg=' and name 'A name for this job' has been
launched on the cluster 'Kubernetes'.The function schedule_workbench_job() schedules a Workbench Job to launch on the default Kubernetes cluster at a specific time and/or at regular intervals.
schedule_workbench_job(
job_name,
schedule_name,
due,
rpt,
project_path,
script,
n_cpu,
n_ram
)A descriptive name to assign to the Workbench Job. Default is the filename of script.
A name, conforming to R variable naming conventions, of the event loop to schedule the Workbench Job on. If the event loop does not already exist, it will be created.
The date and time (as a POSIXct object) to run the Workbench Job.
Run the Workbench Job every rpt seconds after the first scheduled run. For example, to run once every day, this value should be 86,400 seconds. Defaults to NULL and the Workbench Job will only be run once at the date and time specified in due.
The full file path to the root directory of the project that script should be run in e.g. /path/to/project/.
A file path to the R script to be run by the Workbench Job, relative to project path e.g. code/script.R.
The number of CPUs to request for the Workbench Job e.g. 1.0. Default is the default number of CPUs for a session as defined in the Kubernetes cluster.
The amount of memory (RAM) in MiB to request for the Workbench Job e.g. 4096. Default is the default amount of memory for a session as defined in the Kubernetes cluster.
schedule_workbench_job(
job_name = "job_name_1",
schedule_name = "schedule_1",
due = lubridate::now() + lubridate::seconds(10),
rpt = as.numeric(lubridate::as.duration("3 days")),
project_path = here::here(),
script = "code/test_launch_workbench_job_script.R",
n_cpu = 0.25,
n_ram = 128
)
ℹ The schedule 'schedule_1' will submit a Workbench Job with the name 'job_name_1' to be launched on Friday 13 January
2023 at 12:37:33.
✔ A Workbench Job with id 'S3ViZXJuZXRlczpqb2ItbmFtZS0xLXg4bW1j' and name 'job_name_1' has been launched on the cluster
'Kubernetes'.
ℹ The schedule 'schedule_1' will submit a Workbench Job with the name 'job_name_1' to be launched on Monday 16 January
2023 at 12:37:33.A specific scheduled Workbench Job cannot be cancelled. Instead, please refer to cancelling a schedule
The function cancel_workbench_job_schedule() cancels a schedule created by the function schedule_workbench_job(). Cancelling a schedule destroys the event loop that runs the schedule, then removes the object from R's Global Environment.
cancel_workbench_job_schedule(
schedule_name
)The name (as a string) of the schedule (event loop object) to cancel.
cancel_workbench_job_schedule("schedule_1")
✔ The schedule 'schedule_1' has successfully been cancelled.
ℹ No further Workbench Jobs that were scheduled on 'schedule_1' will be launched.
ℹ The schedule 'schedule_2' is still active.