Call this inside a stage function to write a structured message to the
run's event log. The message appears in pipeline_top() event feeds,
failure_timeline(), and the raw JSONL event store. It never throws
an error – if logging fails (e.g., outside a pipeline run), the call
is silently ignored.
Usage
parade_log(msg, severity = "info", ...)
Arguments
- msg
Character message to log
- severity
One of "info" (default), "warn", or "error"
- ...
Additional named fields to include in the event record
(e.g., iteration = 5L, metric = 0.83)
Value
NULL (invisible), called for side effect
Examples
# \donttest{
fl <- flow(data.frame(x = 1:4)) |>
stage("compute", function(x) {
parade_log("starting heavy computation", iteration = x)
result <- x^2
if (result > 10) parade_log("large result detected", severity = "warn")
list(y = result)
}, schema = returns(y = dbl()))
# }