THIS MODULE IS EXPERIMENTAL AND SUBJECT TO CHANGE

Helper functions to provide visibility into runtime errors within a running Bandit instance

Can be used within an IEx session attached to a running Bandit instance, as follows:

iex> Bandit.Trace.start_tracing()
... # Wait for traces to show up whenever exceptions are raised
iex> Bandit.Trace.stop_tracing()copy

It can also be started within your application by adding Bandit.Trace to your process tree.

Bandit.Trace will emit a trace on every exception that Bandit sees (both those emitted from within your Plug as well as internal ones due to protocol violations and the like). These traces consist of a complete dump of all telemetry events that occur in the offending request's parent connection.

Tracing imposes a modest but non-zero load; it should be safe to run in most production environments, but it is not intended to run on an ongoing basis.

By default, Bandit.Trace maintains a FIFO log of the last 10000 telemetry events that Bandit has emitted. Events which correlate to the parent connection which have been evicted from this queue will not be included in this output.

WARNING The emitted logs contains a complete copy of your request's Plug data, as well as all data sent and received on all requests which are contained in the output. It is therefore of the utmost importance that you carefully redact the output before sharing it publicly.