2.1.0

1.Support server performance monitoring (such as qps and rt) based on prometheus.
2.Support server online/offline operation. (Mainly used for online troubleshooting).

2.0.2

1.unify service discovery logic for dashboard and client
2.support registration server to multiple registration centers
3.actively clear registration information when the server is shut down
4.support configuring the maximum number of retries and maximum retry traffic for each domain
5.support setting delayed cancelation based on exception type.
6.add id field to uniquely identify transaction, used during updates and deletes.
7.use a separate thread pool to handle clients' REGISTER commands to avoid block other commands
8.optimize alarm threshold configuration
9.optimize channel select strategic during recovery

release 2.0.1

1. add MaxTimeTryingTreatAsFailed config
2. retain at least one channel from client to server even if no traffic.

2.0.0

Merge pull request #386 from nervose/master-2.x-local-dev

add netty length field length

1.8.1

fix dubbo online npe issue.

1.8.0

1. support multiple tier tcc app.

1.7.0

Merge branch 'refactoring-1.6.x' into master-1.6.x

* refactoring-1.6.x:
  update the tcc transaction dubbo, remove tccJavassist and tccJdk.
  update the TransactionContextEditor
  fix sample error.
  add CompensableTransactionFilter, remove TccJavassist.
  add dubbo filter

# Conflicts:
#	tcc-transaction-dubbo/src/main/java/org/mengyun/tcctransaction/dubbo/filter/CompensableTransactionFilter.java
#	tcc-transaction-dubbo/src/main/resources/META-INF/dubbo/org.apache.dubbo.rpc.Filter

1.6.3

this is latest production line.