Atomstar's blog

Goal: set up data downsampling scheme on existing collectd data.

This article elaborates the official InfluxDB guide here (https://docs.influxdata.c...wnsampling_and_retention/)
where we already have a database with existing measurements that we want to
keep.

I'd like to downsample my collectd data on ping, memory, load, etc. data from
two machines. I only want high (minute) time resolution for a few days, then
downsample to save space.

Update:

• instead of using different field/measurement names (e.g. mean_value in ping_hourly), one can use the same measurement and field name to make querying more easy.
• using grafana variables, one can use dynamic retention policy (RP) use, e.g. use the right RP for the right time range

Read more »

When we were looking to buy a house we did not use a purchase broker (aankoopmakelaar), but instead looked for houses ourselves online (mostly via funda.nl). One thing we often heard was that we’d miss houses were sold silently, from broker to broker. To verify this rumour, I bought kadaster data of recently sold houses in our neighbourhood and checked whether I could find these online.

Findings:

1. 18 out of 20 houses had advertisements online
2. 1. 10 on funda
   2. 8 only on local brokers
3. 2 houses were not advertised publicly, where one was sold at a very low price (perhaps sold to relatives?)

Method:

1. Buy a few zip code reports from the kadaster
2. For each house sold recently, search on google/funda

Conclusion: 90% of houses for sale are available online, of which at least 50% on Funda (some might have expired), and 40% on local websites. The hidden market could be limited to 10%.

Update 20191020: replies disabled because of spambots

Read more »