Gather metrics in production

[vmarkovtsev]

• Write abstract code to submit metrics
• Wait for https://github.com/src-d/infrastructure/issues/269
• Write the code to integrate with (2)

[zurk]

which metric do you mean?

[vmarkovtsev]

Production metrics. They are usually plotted on dashboards. The difference to logs is that you inspect graphs and see the dynamics.

[zurk]

So, about time, system load, etc?

[zurk]

ok, I think I get an answer from the attached infra issue. 👍

[vmarkovtsev]

Time and system load are a tiny fraction of the metrics we will collect. I am talking about production ML dashboards with hundreds of graphs. Example of metrics in the core: src-d/lookout-sdk-ml@2767f72