Open elastic-apm-tech opened 1 year ago
@AlexanderWert Does https://github.com/elastic/kibana/pull/157724/files#diff-b0f9281b29e17f3aa9406728abcfb41a649247dcf8ea1fb0dc801ad4439cd783 already provide a dashboard for Node.js runtime metrics? That "nodejs.json" file includes \"sourceField\":\"nodejs.memory.heap.used.bytes\"
, for example.
Granted this doesn't handle any OTel SDK possible metrics.
Does https://github.com/elastic/kibana/pull/157724/files#diff-b0f9281b29e17f3aa9406728abcfb41a649247dcf8ea1fb0dc801ad4439cd783 already provide a dashboard for Node.js runtime metrics?
Answering my own question: Yes, it does. That dashboard is clearly there in my local-dev 8.10 kibana.
Alex, please correct me if I'm wrong. Our goal for this issue for Node.js then is to:
agentName === 'opentelemetry/nodejs'
for default (or common) Node.js runtime metrics provided by the OTel JS SDK; and(/cc @david-luna)
Found the host-metrics
Otel package is sending bad measurements for CPU utilization. Filed an issue
https://github.com/open-telemetry/opentelemetry-js-contrib/issues/1718
PR on apm-contrib
repo to add memory metrics which could be reused in other languages
Alignment in process.cpu and system.cpu metrics at semconv level has been suggested https://github.com/open-telemetry/semantic-conventions/issues/563
PR to fix process.cpu.time
& process.cpu.utilization
https://github.com/open-telemetry/opentelemetry-js-contrib/pull/1785
There is a PR to update process semconv which may change the attribute names. The PR above is prepared to receive the necessary changes
See meta issue for the description and details: