I'm using influxdb to store values from Iobroker. However I have the issue that the DB is growing much faster than the content I'm storing in it. How can I stop it from growing (without reducing the raw data I want to store or the retention policy)? (Current retention policy is forever for the bucket I'm storing the data in)
As a measure of what creates the big amount of data, I checked what "_measurements" are written during a day to the database. I found that the majority of rows is created not by the data I want to store but by something else. Below are some of measurements with most datapoints - It seems they are coming from Influxdb - but do I need them or can I somehow get rid of them?
| _measurement | counts per day |
|---|---|
| http_api_request_duration_seconds | 3850854 |
| service_bucket_new_duration | 604800 |
| storage_compactions_duration_seconds | 567096 |
| query_influxdb_source_read_request_duration_seconds | 364050 |
I had the same problem and it took me awhile to figure it out.
In the Influx UI check if you have a Scraper setup. The URL for the Scrapers page is something like this:
http://localhost:8086/orgs/YOUR_ORG/load-data/scrapers
On that page I had a scraper with URL
http://localhost:8086/metricsconfigured with my custom bucket. Changing that Scraper to another bucket with a short retention policy resolved the issue.