Prisma telemetry operation contains unaccounted for time
Hi there! I'm looking for a little help understanding something I'm seeing frequently from opentelmetry traces in my application.
https://www.prisma.io/docs/orm/prisma-client/observability-and-logging/opentelemetry-tracing#trace-output
I'm seeing many prisma:client:operation spans with unaccounted for time in the operation itself. I see serialization, the db query itself, and deserialization, but then there's over 600ms of unaccounted for time remaining in the operation. Where could this be coming from? Is it possible to get more detail out of this? I've attached a sample tree of spans in my trace that illustrates.
https://cdn.discordapp.com/attachments/937751383124361286/1245834663285489825/image.png?ex=665a3172&is=6658dff2&hm=e2afee95d2aa71e5eb4e3e5d3421f3b086831bae1e6d46f31d83f6dcda37b51d&
1 Reply
Trace visualization is created using jaeger. Prisma instrumentation is instantiated and registered with the config:
new PrismaInstrumentation({ middleware: true })