Trace span metrics

[kailash8465]

i need help to understand how the following metrics will be useful
traces_spanmetrics_calls_total
traces_spanmetrics_latency_bucket
traces_spanmetrics_latency_count
traces_spanmetrics_latency_sum
traces_spanmetrics_size_total

[mariorodriguez]

Hi @kailash8465, spanmetrics can be useful in multiple different ways. Here are some of them I can think of, but surely not all:

• If your systems are not instrumented with metrics, you get metrics ‘out-of-the-box’ from your traces
• They can provide in-depth monitoring, as generally tracing instrumentation reaches deeper into an applications code.
• They are automatically generated with exemplars. Exemplars are an exceptional way of finding outliers.
• Spanmetrics provide visibility on which services or tenants are sending the most traces to your tracing backend.
• Can help you understand how a trace/span compares to other traces, in terms of latency, size, etc.

In summary, spanmetrics provides an aggregated view of your tracing data. While the general purpose of tracing is providing a detailed view of a request (or other) within a system, sometimes you need a broader picture to fully understand that trace. That broader picture can be provided by spanmetrics.