Details
-
New Feature
-
Status: Open
-
Major
-
Resolution: Unresolved
-
None
-
None
Description
In large-scale production environment across cross datacenters, there may be few clusters (for example, HDFS storage) collecting data from apps in multiple datacenters. For security reasons, most of the time flow goes through load-balancer and Ops/Admin requires visibility of RPC latency across WAN.
Avro IPC supports trace collection [1], it would be good if we make it as configurable inside flume and stores metrics for further analysis.
Attachments
Issue Links
- relates to
-
FLUME-748 Create metric collection infrastructure
-
- Closed
-