You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: A87-grpc-metrics-xds-outlier-detection.md
+6-6Lines changed: 6 additions & 6 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -12,18 +12,18 @@ This document proposes some new metrics that will be added in gRPC for xDS Clien
12
12
13
13
## Background
14
14
15
-
A50: gRPC xDS Outlier Detection Support is a spec for gRPC to support xDS Outlier Detection. The current implementation only offers debug and trace logging in terms of visibility, which can be insufficient for understanding and diagnosing decision making in large-scale production systems.
15
+
[A50: gRPC xDS Outlier Detection Support](https://github.com/grpc/proposal/blob/master/A50-xds-outlier-detection.md) is a spec for gRPC to support xDS Outlier Detection. The current implementation only offers debug and trace logging in terms of visibility, which can be insufficient for understanding and diagnosing decision making in large-scale production systems.
16
16
17
17
Using A79: Non-per-call Metrics Architecture, it is possible to add granular metrics to make visibility into outlier detection easy for service owners utilizing gRPC.
A79’s non-per-call metrics architecture fits perfectly into the metrics reporting solution. We propose to aim for parity where appropriate with Envoy’s xDS implementation for metrics to be collected in gRPC
26
+
[A79’s](https://github.com/grpc/proposal/blob/master/A79-non-per-call-metrics-architecture.md) non-per-call metrics architecture fits perfectly into the metrics reporting solution. We propose to aim for parity where appropriate with [Envoy’s Outlier Detection metric collection](https://www.envoyproxy.io/docs/envoy/latest/configuration/upstream/cluster_manager/cluster_stats#outlier-detection-statistics) to be collected in gRPC
27
27
28
28
Outlier Detection metrics will have the following labels:
29
29
@@ -35,7 +35,7 @@ The following metrics will be exported:
| grpc.xds_client.outlier_detection.ejections_enforced_total | Counter | {ejection} | grpc.target Total enforced ejections due to any outlier type |
38
+
| grpc.xds_client.outlier_detection.ejections_enforced_total | Counter | {ejection} | grpc.target| Total enforced ejections due to any outlier type |
39
39
| grpc.xds_client.outlier_detection.ejections_active | Gauge | {ejection} | grpc.target | Number of currently ejected hosts |
40
40
| grpc.xds_client.outlier_detection.ejections_overflow | Counter | {ejection} | grpc.target | Number of ejections aborted due to max ejection percentage |
0 commit comments