Pushgateway Integration — push_to_gateway • openmetrics

Some workloads may not want to run an HTTP server to expose metrics, especially in the case of short-lived batch jobs. For these cases metrics can also be manually "pushed" to a Prometheus Pushgateway instance, though there are drawbacks.

push_to_gateway() is used to push metrics, and delete_from_gateway() is used to clean them up when the workload is finished.

push_to_gateway(
  url,
  job,
  instance = NA,
  registry = global_registry(),
  ...,
  retry_times = 3,
  httr_config = list()
)

delete_from_gateway(url, job, instance = NA, ...)

Arguments

url	The URL of the Pushgateway
job	A value for the `job` label applied to all pushed metrics.
instance	A value for the `instance` label applied to all pushed metrics, or `NA` to leave it unset.
registry	A `Registry` object, defaulting to the shared global one.
...	Additional named string arguments converted to labels. Beware that these are not yet checked for URL safety.
retry_times	Number of times to (re)try the request if it fails. Passed to `httr::RETRY()`.
httr_config	Additional request configuration (such as timeouts or SSL options) passed to `httr::RETRY()`.

Value

NULL, invisibly.

Examples

if (FALSE) {
register_default_metrics()
push_to_gateway("localhost:9091", job = "batch-job-1")
# Some time later...
delete_from_gateway("localhost:9091", job = "batch-job-1")
}