Just looking for general advice here.
We currently have a system of data recorders, recording pressure and environmental data with particle electrons in the field and we are switching to blues. We originally used MySQL and then migrated to PostgresDB. The database is currently hitting 40GB for 250 devices and growing.
I don’t necessarily need to migrate the data, but as we move to 1,000s of devices, I am sure there are better databases out there.
We currently are looking at something like mongoDB or Redis, hosted on Digital Ocean. We need to archive, retrieve and export customer data, but we also need to graph and generate reports. I have currently created this project in PHP with a javascript layer utilizing highcharts. I was wondering if Grafana (or another product) is a better choice.
Besides time series data, we also need to store customer service and configuration data.
Any ideas or experiences would be helpful
I’ve been using a PostgreSQL database server for going on 2 years now. System add about 40k records per hour. Database size is about 550 GB.
My use is pretty simple, but I am very happy with PostgreSQL. It’s JSON abilities add a very useful layer on top of the solid performance and reliability. I explored Influx with Grafana, but moved back to PostgreSQL with Grafana and Jupyter notebooks for analytics and graphics. Less to learn, larger knowledge/support universe, as it is difficult to not have a relational db as core for IoT data collection. I find using simple triggers to move incoming records to analytic and archive tables reliable.
Timescale seems a solid addition to PostgreSQL to compete with Influx and other for high performance indexing and it seems more opportunity to keep your raw data.
I am sure there is a upper limit on data ingress with PostgreSQL and other relationals vs. the non-relational log and similars. However those seem pretty high and partitioning seems a good practice in most use cases for customer reliability.
That said, there seem to be happy Influx users.
2 Likes