Paul will talk about the long-term vision for Flux the language as well as InfluxDB 2.0, Telegraf 2.0 and beyond. He’ll talk about why we’ve decided to create a language, how that plays into polyglot persistence & purpose-built time series databases, and how it enables more complex analytics and processing workloads to drive insights from data not just in InfluxDB, but everywhere.

1. InfluxDB 2.0, Flux, and Telegraf
Paul Dix
paul@influxdata.com
@pauldix

2. 2.0

5. Why?

6. Optimize for Developer
Happiness

7. Optimize Time to Awesome

8. How?

9. Influx = Vertically Integrated

15. More configuration, more code

16. Convention over configuration

17. No code > code

18. Reusable Components

19. User & Developer Experience

20. Edge
InfluxDB API
Telegraf Client Libs
InfluxDB
OSS
Prometheus Other
InfluxDB UI Grafana
Other
Systems

21. Edge
InfluxDB API
Telegraf Client Libs
InfluxDB
OSS
Prometheus Other
InfluxDB UI Grafana
Other
Systems

22. Edge
InfluxDB API
Telegraf Client Libs
InfluxDB
OSS
Prometheus Other
InfluxDB UI Grafana
Other
Systems

23. Edge
InfluxDB API
Telegraf Client Libs
InfluxDB
OSS
Prometheus Other
InfluxDB UI Grafana
Other
Systems

24. Edge
InfluxDB API
Telegraf Client Libs
InfluxDB
OSS
Prometheus Other
InfluxDB UI Grafana
Other
Systems

26. Personas
• Operator
• User
• Power User
• Developer/Application Builder

27. Setup (operator)
• Sign up for InfluxCloud
• Collect the data
• Telegraf
• Client libraries
• InfluxDB OSS
• Prometheus, Vector, Beats, other
• UI elements to control collection, configuration, etc

28. Users

32. Power Users

38. import "http"
import "json"
option task = {
name: "get-my-data",
every: 1h,
}
response = http.Get(
url: "https://s3.amazonaws.com/influxdata-sample/test-data.json")
result = json.Parse(response.body)
jsonToTable(result.object) |> to(bucket: "mydata")

39. Developers

40. Types of development
• Flux scripts
• Tasks
• Flux packages
• In-platform applications
• Applications that use the platform

41. Environment
InfluxDB API
InfluxDB
OSSMy laptop Influx CLI Flux My App

42. Developing Flux Scripts

43. // get all data from the telegraf db
from(bucket:”telegraf/autogen”)
// filter that by the last hour
|> range(start:-1h)
// filter further by series with a specific measurement and field
|> filter(fn: (r) => r._measurement == "cpu" and r._field == "usage_system")

44. Syntax completion

45. f…

46. from(bucket:”█”)

47. Data completion

48. from(bucket:”█”)

49. from(bucket:”te…”)

50. from(bucket:”telegraf/autogen”)
|>

51. from(bucket:”telegraf/autogen”)
|> r…

52. from(bucket:”telegraf/autogen”)
|> range(start:-1h)

53. from(bucket:”telegraf/autogen”)
|> range(start:-1h)
|> f…

54. from(bucket:”telegraf/autogen”)
|> range(start:-1h)
|> filter(fn: (r) => r.)

55. from(bucket:”telegraf/autogen”)
|> range(start:-1h)
|> filter(fn: (r) => r._measurement )

56. from(bucket:”telegraf/autogen”)
|> range(start:-1h)
|> filter(fn: (r) => r._measurement == "█")

57. from(bucket:”telegraf/autogen”)
|> range(start:-1h)
|> filter(fn: (r) => r._measurement == "cpu" and r._field == "usage_system")

58. Iteratively build and look at
data

59. Developing Tasks

60. CLI as test runner

61. CLI to pull production/test data

62. CLI to push to production

65. Developing Flux Packages

66. influx package init

67. option package = {
name: "foo", // required, must match the name of the package declaration above
description: "an example package", // required
authors: [“Paul Dix <paul@influxdata.com>”], // required, at least 1
version: "0.1.0", // required
license: "MIT", // required
homepage: "https://foo.com", // optional
documentation: "https://foo.com/docs", // optional
repository: "https://github.com/pauldix/foo", //optional
flux_versions: ["0.*.*"], // optional, versions of flux this works with
tags: ["example"], // optional
files: [
"README.md",
"stuff.flux",
"other.flux",
], // optional, all code can be contained in package.flux file
tests: [
"stuff_test.flux",
], // optional
dependencies: [
{package: "nathaniel/bar", version:”>=1.2.0”}
], // only required for any other package repo dependencies
}

68. option package = {
name: "foo", // required, must match the name of the package declaration above
description: "an example package", // required
authors: [“Paul Dix <paul@influxdata.com>”], // required, at least 1
version: "0.1.0", // required
license: "MIT", // required
homepage: "https://foo.com", // optional
documentation: "https://foo.com/docs", // optional
repository: "https://github.com/pauldix/foo", //optional
flux_versions: ["0.*.*"], // optional, versions of flux this works with
tags: ["example"], // optional
files: [
"README.md",
"stuff.flux",
"other.flux",
], // optional, all code can be contained in package.flux file
tests: [
"stuff_test.flux",
], // optional
dependencies: [
{package: "nathaniel/bar", version: “>=1.2.0”}
], // only required for any other package repo dependencies
}

69. influx package test

70. influx package publish

71. import "/pauldix/foo", “>=0.1.1”

72. import "/pauldix/foo", “0.1.1"
// if $FLUX_PATH/pauldix/foo exists
// 1. check package.flux version
// 2. look in _versions/ for matching
// if not, request matching version from repository

73. import "/pauldix/foo", "0.1.2"
// if $FLUX_PATH/pauldix/foo exists
// 1. check package.flux version
// 2. look in _versions/ for matching
// if not, request matching version from repository
foo.myFunc()

74. Not just for Flux

75. option package = {
name: "foo", // required, must match the name of the package declaration above
description: "an example package", // required
author: [“Paul Dix <paul@influxdata.com>”], // required
version: "0.1.0", // required
license: "MIT", // required
homepage: "https://foo.com", // optional
documentation: "https://foo.com/docs", // optional
repository: "https://github.com/pauldix/foo", //optional
flux_versions: ["0.*.*"], // optional, versions of flux this works with
tags: ["example"], // optional
files: [
"README.md",
"stuff.flux",
"other.flux",
], // optional, all code can be contained in package.flux file
tests: [
"stuff_test.flux",
], // optional
dependencies: [
{package: "nathaniel/bar", version: “>=1.2.0”}
], // only required for any other package repo dependencies
}

76. option package = {
name: "averages",
description: "an example downsampling task package",
authors: [“Paul Dix <paul@influxdata.com>”],
version: "0.1.0",
license: "MIT",
type: "task", // could be flux, or application
files: [
"README.md",
"downsample.flux",
],
}

77. Developing Custom
Applications

78. InfluxDB OSS as local dev
environment

79. https://github.com/influxdata/giraffe

80. https://github.com/influxdata/clockface

81. Better Developer Ergonomics =
Increased Productivity

82. What’s Available Today
• Cloud 2
• OSS Alpha 18
• Monitoring and Alerting
• Pager Duty
• Slack
• HTTP POST
• Flux
• Superset of InfluxQL
• String interpolation
• Dynamic Queries

83. Getting to InfluxDB 2.0 OSS GA
• InfluxDB 1.x comparability mode (writes, InfluxQL)
• Upgrade migration
• TLS support
• Finalizing Flux
• Error handling
• Filter based on field type
• Windowing by calendar month
• Backup/restore or bulk import/export
• Beta -> Performance, bug fix -> GA

84. Coming Soon
• Flux
• User Packages
• Recursion & Looping
• HTTP methods & input parsing
• Developer Ergonomics
• OSS integration with Cloud 2

85. Thank you.
Paul Dix
paul@influxdata.com
@pauldix