Introduce pull based replicator that does not use kafka (cadence-work…

…flow#2377) - Created replication message fetcher and processor on standby history side. Fetcher is per node and processor is per shard. - Fetch requests are aggregated on box level and then send to source frontend. - Implemented fan out logic on source frontend to get replication messages from local history shards. - Introduced replication consumer config to control whether to use kafka based consumer or rpc based consumer.
MirrorFork · Aug 15, 2019 · 602b810 · 602b810
1 parent 6c55787
commit 602b810
Show file tree

Hide file tree

Showing 72 changed files with 4,293 additions and 216 deletions.
diff --git a/.gen/go/cadence/cadence.go b/.gen/go/cadence/cadence.go
diff --git a/.gen/go/cadence/workflowserviceclient/client.go b/.gen/go/cadence/workflowserviceclient/client.go
diff --git a/.gen/go/cadence/workflowserviceserver/server.go b/.gen/go/cadence/workflowserviceserver/server.go
diff --git a/.gen/go/cadence/workflowservicetest/client.go b/.gen/go/cadence/workflowservicetest/client.go
diff --git a/.gen/go/history/history.go b/.gen/go/history/history.go
diff --git a/.gen/go/history/historyserviceclient/client.go b/.gen/go/history/historyserviceclient/client.go
diff --git a/.gen/go/history/historyserviceserver/server.go b/.gen/go/history/historyserviceserver/server.go
diff --git a/.gen/go/history/historyservicetest/client.go b/.gen/go/history/historyservicetest/client.go