Time-series data: InfluxDB vs TimescaleDB vs ClickHouse for monitoring

we're building out a new internal monitoring and observability platform, and we need to store a massive amount of time-series data, roughly a million data points per minute. we're currently evaluating a few options. influxdb was an early contender, but we're wary of potential licensing changes or vendor lock-in with its enterprise features. timescaledb is appealing because we're already heavily invested in postgresql, and the familiarity would be a huge plus, but we wonder about its raw performance at this scale. clickhouse is also on the table for its columnar analytics power but represents a new operational burden. for those running high-volume time-series workloads, what have been your experiences with these or other solutions, especially regarding scalability, operational overhead, and querying performance?