Having some issues getting my configuration to work… i’m seeing double events for every conversation.update coming into my python subscriber. the cli is idempotent but it’s processing the same payload twice, causing race conditions.

i’ve tried filtering on the id field in the detail, but the event structure seems to shift slightly between the two pushes. is there a reliable way to dedup in the lambda handler or should i be looking at the EventBridge rule settings?

here’s the handler snippet:

def handler(event, context):

logic here

pass

anyone else hit this? it’s driving me nuts.

Make sure you hash the detail.conversationId combined with detail.timestamp instead of relying on the event id. EventBridge retries often shift metadata slightly, so that composite key stays stable.

here’s the snippet.

import hashlib
dedup_key = hashlib.md5(f"{event['detail']['conversationId']}:{event['detail']['timestamp']}".encode()).hexdigest()
if dedup_key in seen: return