I have been using influxdb as time series database for IoT related data for a long time. But their recent history of product management choices makes me reconsider my position on them for a new software stack for a startup. I wonder what people recommend with the following requirements:
- SQL-like query language
- no limitations on data retention
- allows efficient queries and aggregation over any time interval (group by time interval)
- high industry adoption level, with increased chances of continued long-term service and support
- serverless cloud offering
- not locked in to a specific cloud vendor (multiple cloud vendors available or possibility to run it myself in a VM/containerized solution from any cloud vendor)
- community/FOSS edition that can be run locally for personal/experimental IoT projects
- data retention policies allowing down-sampling of data to reduce long term data storage costs while keeping access to historic values and having high temporal resolution values for recent data (example: keep all values for the past month of data, store median/min/max values per hour for the past half year, and per day for the past 2 years)
- Nice to have: ability to keep a backup of all data in some type of cold storage solution for training ML models in the future.
