Your logs show that the CDC events queue has reached its maximum size (10,000), and Debezium’s MongoDB connector is warning that it “cannot acquire buffer lock” because the buffer queue is full:

CDC events queue stats: size=10000, cap=0, puts=14, polls=0
... buffer queue is likely full

From the available documentation, this queue is the internal CDC buffer that Airbyte uses to hold change events before they’re processed. The “Queue Size” parameter controls how many CDC records can be queued in memory; the recommended default is 10,000. Larger queues use more memory but handle bursts better; smaller queues reduce memory but can hurt sync efficiency. [CDC best practices] The knowledge sources do not provide a specific remediation for this exact “buffer queue is likely full” warning, beyond the general guidance on queue size and its trade‑offs. They also don’t document a MongoDB‑specific setting to change this behavior. Given what is documented, the actionable points you can derive are: • The queue is saturated (size=10000, cap=0), so events are being produced faster than Airbyte can consume them. [CDC best practices] • Changing the queue size is possible but should only be done with a clear technical reason and awareness of memory impact. [CDC best practices] Because the knowledge sources don’t go further, I can’t reliably recommend a specific new queue size or other tuning steps from documentation alone. If this warning is causing sync failures or stalls, the next step would be to: • Collect a full job log around this period. • Open or update a GitHub issue with those logs so the Airbyte team can advise whether to adjust the queue size or investigate a performance bottleneck, similar to how other CDC/queue issues are handled in the referenced GitHub issues. 👍 0 button 👎 0 button ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.