Notes
A new version of the agent has been released. Follow standard procedures to update the Infrastructure agent.
Added
- #155 The Infrastructure agent now handles entities' registration errors the following way (Infrastructure SDK v4 protocol):
- For request errors:
- Connectivity failure or limit exceeded (429 or 503 status codes) will trigger a retry with an exponential backoff with a maximum backoff of 1 minute.
- Other types of errors (including bad payload format) will be logged, and data will be discarded.
- For successful requests (200 status code) containing:
- Errors: The agent will discard related data.
- Warnings will be logged.
- For request errors:
- #156 The Infrastructure agent entity registration now batches requests with a maximum of 1000 entities and payloads of 1Mb per call.
- #123 The embedded telemetry API now splits DM submission by number of entities following requested driven contract limitations. For more informations on the limits, see our docs
Changed
- #206 The built-in Prometheus integration has been updated to version [2.3.0] (https://github.com/newrelic/nri-prometheus/releases/tag/v2.3.0). For more information about the Prometheus integration, see our documentation.
Fixed
- #203 NFS invalid stats fix.
- #188 Prometheus summary type calculated as delta value to properly allow average calculation in NRQL.
- #167 Container processes now correctly filtered as host processes when Process Monitoring is enabled and "include_matching_metrics" rules are defined. For more information about enabling process monitoring and filtering processes, see our documentation.