An_introduction_to_Flux.pdf

Transcript

1. Flux Time series Belgium meet-up
2. New functional language - and engine - to gather and process time-series data Under MIT license Flux
3. Useable Readable Composable Flux: High Level Design Principles (1/2)
4. Testable Contributable Shareable Flux: High Level Design Principles (2/2)
5. Aren't there enough alternatives already?
6. True-SQL vs like-SQL SQL was designed for relational models and calculations Time-series databases are based on a continuous stream of data
7. Like-SQL shortcomings Missing functionalities are frustrating Functional composition is cumbersome
8. Custom scripts and application logic TICK scripts in Kapacitor Python scripts GO daemons
9. Flux language characteristics (1/8) Gather and process time-series data - without changing its state Database agnostic Flow-based functional processing
10. Flux language characteristics (2/8) from(bucket: "MySensorData") |> filter(fn: (r) => r["_measurement"] == "TempSensor1") |> range(start: -30m) |> window(every:10m) |> mean() |> filter(fn: (r) => r._value > 20)
11. Flux language characteristics (3/8) Piping forward from(bucket: "MySensorData") |> filter(fn: (r) => r["_measurement"] == "TempSensor1") |> range(start: -30m) |> window(every:10m) |> mean() |> filter(fn: (r) => r._value > 20)
12. Flux language characteristics (4/8) Functional from(bucket: "MySensorData") |> filter(fn: (r) => r["_measurement"] == "TempSensor1") |> range(start: -30m) |> window(every:10m) |> mean() |> filter(fn: (r) => r._value > 20)
13. Flux language characteristics (5/8) Anonymous functions from(bucket: "MySensorData") |> filter(fn: (r) => r["_measurement"] == "TempSensor1") |> range(start: -30m) |> window(every:10m) |> mean() |> filter(fn: (r) => r._value > 20)
14. Flux language characteristics (6/8) Custom functions square = (table=<-) => { table |> map(fn: (r) => r._value = r._value * r._value) }
15. Flux language characteristics (7/8) SQL HAVING functionality from(bucket: "MySensorData") |> filter(fn: (r) => r["_measurement"] == "TempSensor1") |> range(start: -30m) |> window(every:10m) |> mean() |> filter(fn: (r) => r._value > 20)
16. Flux language characteristics (8/8) SQL JOIN functionality current = from(bucket: "tanksystem") |> filter(fn: (r) => (r.pump == "I1")) |> range(start: -1h) past = from(bucket: "tanksystem") |> filter(fn: (r) => (r.pump == "I1")) |> range(start: -2h, stop: -1h) |> shift(shift:1h) difference = join(on: ["_time"], tables: {current:current,past:past}) |> map(fn: (r) => r.current__value - r.past__value) |> limit(n:15)
17. Demo
18. Release dates Full release coming fall with Influx 2.0 Spec https://github.com/influxdata/flux/blob/m aster/docs/SPEC.md Grafana plugin https://grafana.com/plugins/grafana-influx db-flux-datasource/installation
19. FACTRY Open Manufacturing Intelligence Yves Bourgeois yves.bourgeois@factry.io