We built a "Genesys Bot Connector" integration into our platform. The setup is a "web messenger" using an "inbound message flow" forwarding to our endpoint via a "call bot connector" node.
Normally, we're receiving user inputs as requests to our endpoint and our responses are rendered in the web messenger as incoming messages.
However, if the user sends messages "too fast", our endpoint stops receiving user inputs from the bot connector and we can't react to user messages for about 5 minutes.
As an example: we have an integration which replies to user messages by quoting their message payload.
If the user sends "A", the response will contain one message with text "You said A".
If we send a single message in the web messenger, we're getting a normal response:
User: "A"
Bot: "You said A"
If we send multiple user messages in a fast sequence, we're ending up missing messages:
User: "1"
User: "2"
User: "3"
Bot: "You said 1"
User: "4"
User: "5"
five minutes later
User: "6"
Bot: "You said 4, 5, 6"
Expected:
we will receive a total of 5 HTTP requests from the bot connector, one for each message
Actual:
We're receiving only some of the user inputs as HTTP requests. We checked whether they are delayed, but never received them.
If the user then proceeds to send more messages afterwards, none of them are forwarded to our platform for about 5 minutes.
After waiting for about 5 minutes, when we send another user message through the web messenger, we will receive a concatenated text input containing some, but not all of the missing messages in our platform.
When we observe the conversation history in the agent UI after forwarding the conversation into a queue, all user messages, including those that weren't forwarded, are shown in the conversation history, without concatenation.
Can you please provide the genesys cloud region and orgIdwhere you have run this test?
And if you can also possibly provide the conversationId of this web messaging conversation?
Organization Id: 6db82575-3d4a-4778-8158-bd4478dad1f8
Organization region: eu-central
Conversation Id of session 1: 394a473d-5f7e-495b-867f-bf3d719c7e7a
Conversation Id of session 2: 1c786ed6-d24a-4f9f-9e2e-5d39d4b39e81
@dilek_imseytoglu You have posted the same conversationIds in your two posts.
Which one is related to the random error question and which one to the multiple input question?
@Jerome.Saint-Marc we are facing these issues in almost every conversation, thats why we have added the same convIds to both Especially the 2. Conv Id shows the multiple input issue.
@dilek_imseytoglu My comment was that, as you seem to be able to replicate these issues in your environment, it would make it easier for people who will troubleshoot to have a conversation with problem 1 and a conversation with problem 2. If you share conversationIds without saying what problem it is supposed to include, it doesn't help. I will mention to engineering that the second convId shows the multiple input issue.
@Jerome.Saint-Marc do we have an update on this?
As mentioned in the other topic, I'm unable to recreate the scenario because I can't login to genesys right now.
The "nomatcherror" issue from the other thread is also popping up occasionally for every message, making it hard to have a standalone conversation for this issue since any of the messages could result in a cancelled session on top.
If you can help me log in to the genesys UI, I'm happy to demonstrate the issues to you.
For your Genesys Cloud Web app issue, just make sure to logout, close your browser and clean cookies, and login again.
I don't have update from engineering. The person is out of office this week.
What I would recommend is to open a SERVOPS ticket (the Developer Forum is not a support site with tickets and SLAs) for each of this issue. Send me the ticket number and I will forward it to the engineer.
You now have access to SERVOPS (Ankita requested some access/accounts last week).
Especially the 2. Conv Id shows the multiple input issue.