A lot gets said about how data can help us with observability, but much less gets said about how much we can shape the collection itself to actually help us make decisions.
A few months ago, here at Appoena, I ran into exactly that kind of challenge: a client's team needed to closely track the metrics from their K6 load tests in Datadog, alongside the rest of their observability stack.
The problem
Unfortunately, K6 v0.55.0 dropped support for sending metrics via StatsD, which happens to be Datadog's recommended way of receiving them. We looked into other ways of getting K6 metrics into Datadog, but every alternative came with trade-offs, including the absence of some extremely important p95 metrics.
Digging into the docs
