project
Telemetry pipeline component described as being run per region or per cluster, with batch and memory limiter tuning and bottleneck risk under load.
No evidence-backed relationships are recorded.
Atlassian rebuilt a metrics platform fed by about 100,000 hosts on OpenTelemetry while its applications kept sending StatsD packets to the same address. The platform team did the work alone. Moving thousands of services onto OpenTelemetry SDKs comes later.
Reality
Atlassian's platform team rebuilt collection, ingest, aggregation and forwarding on OpenTelemetry Collectors across 14 regions. Applications kept firing StatsD over UDP, so no service team had to re-instrument first.
Reality
A CNCF walkthrough routes workflow_run and workflow_job events into an OpenTelemetry Collector, so CI coverage depends on org admin access rather than on who remembered to add a tracing step.
Reality
The plugin that used to exist only in Rider now lands in IntelliJ IDEA, GoLand, PyCharm and WebStorm. It removes the local observability backend from the setup cost, not the SDK work.
Reality
Compaction failed sixty times an hour on a full volume while dashboards stayed correct, because queries are served from memory. Liveness checks cannot see durability.
Reality
A devops.com argument for handing off the telemetry pipeline is vendor-adjacent and mostly correct: instrumentation is the cheap part, and the recurring work never appears in a headcount plan.
Reality