🐕🔎 Inspecting big Datadog traces in the CLI
The Datadog trace-viewing UI performance greatly degrades as the number of spans in a single trace increases and within a trace, it is often painful to find the span you are interested in as there is no way to "jump" to a span directly.
This is an attempt at solving that problem: using doggo, you can inspect large trace json blobs and get to the subtree
corresponding to spans of interest in a few keystrokes without having to worry about your browser running out of memory.
At present, you will have to have the json you want to inspect handy -- the tool simply makes it more palatable.
Each release includes pre-built binaries, drop those in your path (i.e. in /usr/local/bin) and you are ready to go!
You can also clone this repository and go build . to get your own executable.
To get doggo on your machine, you can grab a pre-compiled release for your machine if one is available or build it locally by cloning the repository and running go build.
You can either include the executable in your $PATH or run the executable wherever it lives:
doggo <path-to-trace-json> <resource-name> [-vv] [--depth={depth-limit}]
The resource-name provided is the resource name associated with the spans that interest you in the provided trace.
Doggo will display all the span subtrees up to depth-limit depth (unlimited depth if not specified) that have a
resource name that matches your query (either complete or partial).
Default verbosity will include resource names and span duration in millis. Adding -v will toss in some extra
information about each resource and -vv will include any metadata (i.e. tagging) that is available in the trace for
that span.
Doggo welcomes contributions. The main use case is pretty tailored to my day-to-day, but sensible suggested changes are welcome!