config: change log level to debug#3738
Conversation
Signed-off-by: Rama <rama.rao@salesforce.com>
| if (e == nullptr) { | ||
| stats_.update_failure_.inc(); | ||
| ENVOY_LOG(warn, "gRPC update for {} failed", type_url_); | ||
| ENVOY_LOG(debug, "gRPC update for {} failed", type_url_); |
There was a problem hiding this comment.
Can you distinguish between the very first update and alter updates? In particular, I think we should warn when a management server is unreachable etc. but not during normal disconnects (e.g. gRPC load shed).
There was a problem hiding this comment.
If it is not unreachable during Envoy startup, we get the following warning already
[2018-06-27 13:47:24.884][27][warning][upstream] source/common/config/grpc_mux_impl.cc:235] gRPC config stream closed: 14, upstream connect error or disconnect/reset before headers
Is this sufficient? or are you thinking of some thing else?
If we do not use ADS, individual xDS apis may be still logging this for first unreachability also?
There was a problem hiding this comment.
Ok. So do you want this to be warn till at least one successful update has happened or one update attempt? Based on that, I think we can evaluate the corresponding stat and distinguish between debug and warn.
There was a problem hiding this comment.
OK, seems fine then if we are already observing a connection level warning.
2 participants
Signed-off-by: Rama rama.rao@salesforce.com
Description: Changes the log level to debug when management server is not reachable.
Risk Level: Low
Testing: Current tests
Docs Changes: None
Release Notes: NA
Fixes #3733