Hello everyone, We're experiencing a strange issue while using the SigNoz OpenTelemetry (OTel) Collector. Our setup involves: • An OTel Collector running on AWS VMs that sends telemetry data to Kafka. • A SigNoz OTel Collector that reads data from the Kafka cluster and forwards it to ClickHouse. • A self-managed ClickHouse cluster with 3 shards and 3 replicas, with the necessary SigNoz migrations already applied. The issue: Data ingestion into ClickHouse is inconsistent. After restarting the SigNoz collector or modifying its configuration, a small amount of data gets ingested—but it stops shortly afterward. This pattern repeats every time we restart the collector. We’ve enabled debug logs, but they don’t reveal any obvious errors or warnings. Has anyone successfully set up SigNoz with this architecture? Is there something we might be missing—possibly around buffering, batching, or ClickHouse-specific configurations? Any insights would be appreciated!