ack() delayed due to network latency causing event repetition in async Bolt app

[PrajwalNaik07]

Description

When using the async Bolt for Python SDK (slack_bolt==1.23.0), I noticed that calling await ack() inside an event handler (e.g., app_mention) sometimes takes around 3 seconds due to network delays. This delay causes Slack to resend the event multiple times because it does not receive the acknowledgment promptly.

Reproduction Steps

1. Set up an async Bolt app with event handler for app_mention.
2. Call await ack() inside the handler.
3. Deploy the app.
4. Mention the bot in Slack.
5. Observe that if await ack() is delayed (~3 seconds), Slack retries the event, causing duplicates.

Expected behavior

ack() should respond immediately to Slack’s event request to prevent retries, even if the rest of the event processing takes longer.

Actual behavior

await ack() delays due to network latency and event processing, causing Slack to resend the event multiple times.

Code snippet

@app.event("app_mention")
async def bot_mention(event, body, say, client, logger, ack):
    await ack()  # This sometimes takes ~3 seconds due to network delay
    # Additional event processing...

Environment

slack_bolt==1.23.0
slack_sdk==3.36.0
Bolt for Python (async)
Slack workspace with necessary permissions and subscriptions
Python 3.8+

[zimeg]

Hey @PrajwalNaik07! 👋

I understand the 3 second ack timeout can be problematic for certain network conditions, but this seems to be expected behavior to me. The event handler should have automatic acknowledgement too unless other App settings - such as the option process_before_response - are being used:

• We wait longer than 3 seconds to receive a valid response from your server.

🔗 https://docs.slack.dev/apis/events-api/#failure
🔗 https://docs.slack.dev/tools/bolt-python/concepts/event-listening

Other than network latency I am curious if this app is long running or if prior computation is done before the event can be processed? We recommend sending an acknowledgment back as soon as possible to avoid these retries 👾