switch to auto consumer offset updates

This is the classic/correct way to do consumer group updates for higher
throughput, when "at least once" semantics are acceptible (as they are
here; double processing should be safe/fine).

2 files changed, 11 insertions, 2 deletions

diff --git a/python/fatcat_tools/workers/changelog.py b/python/fatcat_tools/workers/changelog.py
index 92bb8bdd..6319da2f 100644
--- a/python/fatcat_tools/workers/changelog.py
+++ b/python/fatcat_tools/workers/changelog.py
@@ -102,12 +102,17 @@ class FatcatEntityUpdatesWorker(FatcatWorker):
             managed=True,
             auto_offset_reset=OffsetType.LATEST,
             reset_offset_on_start=False,
+            fetch_message_max_bytes=4000000, # up to ~4MBytes
+            auto_commit_enable=True,
+            auto_commit_interval_ms=30000, # 30 seconds
+            compacted_topic=True,
         )
 
         with release_topic.get_sync_producer() as producer:
             for msg in consumer:
                 cle = json.loads(msg.value.decode('utf-8'))
                 #print(cle)
+                print("processing changelog index {}".format(cle['index']))
                 release_edits = cle['editgroup']['edits']['releases']
                 for re in release_edits:
                     ident = re['ident']
@@ -118,5 +123,5 @@ class FatcatEntityUpdatesWorker(FatcatWorker):
                         partition_key=ident.encode('utf-8'),
                         timestamp=None,
                     )
-                consumer.commit_offsets()
+                #consumer.commit_offsets()
 
diff --git a/python/fatcat_tools/workers/elastic.py b/python/fatcat_tools/workers/elastic.py
index 119a6a26..3a75a1b3 100644
--- a/python/fatcat_tools/workers/elastic.py
+++ b/python/fatcat_tools/workers/elastic.py
@@ -32,6 +32,10 @@ class FatcatElasticReleaseWorker(FatcatWorker):
         consumer = consume_topic.get_balanced_consumer(
             consumer_group=self.consumer_group,
             managed=True,
+            fetch_message_max_bytes=4000000, # up to ~4MBytes
+            auto_commit_enable=True,
+            auto_commit_interval_ms=30000, # 30 seconds
+            compacted_topic=True,
         )
 
         for msg in consumer:
@@ -45,4 +49,4 @@ class FatcatElasticReleaseWorker(FatcatWorker):
             print("Updating document: {}".format(elastic_endpoint))
             resp = requests.post(elastic_endpoint, json=release_elastic_dict(release))
             assert resp.status_code in (200, 201)
-            consumer.commit_offsets()
+            #consumer.commit_offsets()