We wire up services with tracing. We emit logs. All in service of observability.
What’s going right, what’s going wrong. How are things wired.
But we silo the data. We set up exporters, stores, and GUIs. But only in our k8s environments.
Praytel unifies logs and tracing on the command line so your development workflow can leverage both.
Praytel is a transparent CLI wrapper: praytel [flags] -- command. It runs one
command, captures its stdout/stderr and the OTLP traces and logs it exports,
joins them, and prints a single view. There is no daemon, no collector and no
leftover files.
praytel -- go test ./... # append-only stream
praytel --tui -- ./server # live trace/span tree
It's open source: barry-hennessy/praytel on GitHub.
How to run an app under praytel and watch its traces arrive as a live span tree.
How praytel attaches each log line to the span that produced it, and how to make your logs joinable.
How to use the graph view to see which components a run touches and how they call each other.
Praytel v0.1.0 is out: Traces and logs on the command line