trk

Really, really simple terminal-based time-series data recorder and plotter.

I have quite a few scripts which inject data into off-the-shelf apps, generate reports, compare data in various systems and more. All the usual fun things you do when you're working with a pile of legacy infrastructure (and cleaning it up).

This helps me track how many issues there are over time and quickly and easily identify when large changes happen.

Usage

```bash

Add a new series with specific units

$ trk add-series -n inv.req.time -u ms $ trk add-series -n devices.alive

Or interactively

$ trk add-series Series Name: services.provisioned Input Unit (eg ms, bps): svc Created services.valid

Then add some data points

$ trk add -s inv.req.time 9 $ trk add -s inv.req.time 8 $ trk add -s inv.req.time 9

Or via the bulk command (for integrating with scripts)

$ cat <

$ cat points.txt | trk bulk

You can always use -c to auto-create the series if you don't want to pre-populate

$ trk add -c -s new.series 20

trk stores files in $HOME/.trk by default, customise this if desired

$ trk -d ~/my-metrics bulk -c < points.txt

Or use different data files to keep things separate

$ trk -f app1 add -c -s metric.a 33 $ trk -f app2 add -c -s metric.a 983

And then plot the output (see screenshots below, the braille text doesn't space correctly here)

$ trk plot -s inv.req.time

If you want the detail, add a table to

$ trk plot -s inv.req.time -t

Or if you're not sure what you've been plotting, don't specify a series and use the interactive list

$ trk plot ```

Contributing

Send a PR.

I'm fairly new using Rust so I expect that there's a lot which could be fixed. Things which are at the top of my (low-priority) todo list:

• Find a nice SQL generator (like Perl's SQL::Abstract or ...)
• Support JSON maybe for bulk ingest?
• Clean up module structure

LICENSE

See LICENSE file.

Code under src/textplots is MIT licensed but (C) Alexey Suslov