[Release] Bump to v0.3.0

CHANGELOG.rst

@@ -2,11 +2,44 @@
 Changelog
 =========
 
-Next release
-============
+Version 0.3.0
+=============
+
+Breaking changes
+****************
+
+- The variables exposed on the metrics endpoint were seriously refactored. The following metrics
+  were removed:
+    * pyaleph_processing_pending_messages_gtasks_total
+    * pyaleph_processing_pending_messages_action_total
+    * pyaleph_processing_pending_messages_messages_actions_total
+    * pyaleph_processing_pending_messages_i_total
+    * pyaleph_processing_pending_messages_j_total
+- The `allow_amend` and `code.comment` fields are now mandatory for PROGRAM messages.
+
+New features
+************
+
+- Users can now query the programs that respond to Aleph messages using the new /programs/on/messages endpoint.
+- Users can now forget aggregates by specifying one or more aggregate keys in the `aggregates` field
+  of the FORGET message.
+- The message processing loop was heavily refactored. Users can now set parallel processing
+  limits by message type to tune the CCN to their network abilities.
+- A lot of improvements were made regarding code quality, by validating user data earlier in the process and using
+  classes after the validation. This results in minor performance improvements and increases maintainability of
+  the code base.
+- We now use rotating log files. This resolves an issue where nodes could go out of disk space because of the size of
+  the log files.
+- Resolved a 500 error that occurred when feeding an invalid hash format to /storage/raw/{hash}.
+
+
+Internal changes
+****************
 
-- Breaking change: the "pyaleph_processing_pending_messages_action_total" and
-  "pyaleph_processing_pending_messages_action_total" values are no longer part of the metrics.
+- A lot more tests!.
+- The throughput of the TXs and message processing loops was improved. The jobs are now guaranteed to execute
+  the same number of tasks over time, instead of a varying amount before.
+- We now use unordered bulk inserts when writing to the DB. This potentially improves performance.
 
 Version 0.2.1
 =============

deployment/samples/docker-compose/docker-compose.yml

@@ -7,7 +7,7 @@ volumes:
 services:
   pyaleph:
     restart: always
-    image: alephim/pyaleph-node:latest
+    image: alephim/pyaleph-node:v0.3.0
     command: --config /opt/pyaleph/config.yml --key-dir /opt/pyaleph/keys -v
     ports:
       - "127.0.0.1:8000:8000/tcp"

deployment/samples/docker-monitoring/docker-compose.yml

@@ -9,7 +9,7 @@ volumes:
 services:
   pyaleph:
     restart: always
-    image: alephim/pyaleph-node:latest
+    image: alephim/pyaleph-node:v0.3.0
     command: --config /opt/pyaleph/config.yml --key-dir /opt/pyaleph/keys -vv
     ports:
       - "127.0.0.1:8000:8000/tcp"

docs/conf.py

@@ -293,5 +293,5 @@
 
 
 rst_epilog = """
-.. |pyaleph_version| replace:: v0.2.1
+.. |pyaleph_version| replace:: v0.3.0
 """

src/aleph/config.py

@@ -17,7 +17,7 @@ def get_defaults():
             "jobs": {
                 "pending_messages": {
                     "max_concurrency": 2000,
-                    "store": 100,
+                    "store": 30,
                 },
                 "pending_txs": {
                     "max_concurrency": 20,