api/filter: add tcp_proxy/rate_limit/mongo_proxy filter config protos.#51
api/filter: add tcp_proxy/rate_limit/mongo_proxy filter config protos.#51htuch merged 2 commits intoenvoyproxy:masterfrom
Conversation
TCP proxy filter now has an idle timeout and the source match (and rest of route match as well) are now in the FilterChainMatch in envoyproxy#49. Fixes #23, #45.
| string stat_prefix = 1; | ||
|
|
||
| // The upstream cluster to connect to. | ||
| string cluster = 2; |
There was a problem hiding this comment.
Is connection-weighted clusters extension to tcp_proxy part of the plan for v2?
We might have some commonality here with RDS weighted clusters.
There was a problem hiding this comment.
I'm wondering how useful fixed weighted balancing at the granularity of TCP connections (vs. requests) is, it probably depends on what's going on in the connections.
There was a problem hiding this comment.
Yes, for long living connections the utility is small. Nevertheless, some application protocols might not have dedicated filters in envoy yet (think smth like redis), where this might be useful.
There was a problem hiding this comment.
Actually, I think WeightedCluster is mostly for canarying, is this right @mattklein123? In that case it seems something useful to have.
There was a problem hiding this comment.
Canary, BG, etc. I think it's worthwhile to have it here, but can definitely be done in follow up since it's an extension to what we have today. We should share the weighted cluster Message with RDS also.
api/filter/tcp_proxy.proto
Outdated
| // filter. The idle timeout is defined as the period in which there is no | ||
| // active traffic. If not set, there is no idle timeout. When the idle timeout | ||
| // is reached the connection will be closed. | ||
| google.protobuf.Duration idle_timeout = 11; |
There was a problem hiding this comment.
Should we split upstream/downstream idle?
There was a problem hiding this comment.
I don't think we need it, but let me know if you'd like it.
There was a problem hiding this comment.
I don't think we need it now, but at minimum I would rename downstream_idle_timeout just so that we can add upstream later. I know this will be asked for at some point.
3 participants
TCP proxy filter now has an idle timeout and the source match (and rest
of route match as well) are now in the FilterChainMatch in
#49.
Fixes #23, #45.