June 2018
Chronograf: Present & Future
Chronograf Goals
Platform
Administration
Data
Exploration
Rapid
Visualization
Sharing Expertise
It’s a Journey…
Where do we go from here?
Data Visualization
Cool!
We can visualize tabular data with Chronograf.
So what??
Log Analysis
• Most common adjunct use case for time series metrics
• Show me what happened around the time of…
• Bring the best of time series to logs
Log Analyzer
Log Analyzer
Filter by…
• Hosts
• Services
• Regex
• Provide Histogram of log output by time
Update Image
Demo
Cool!
Log data visualized with Chronograf.
So what??
Rapid Visualization
&
Problem Isolation
More things…more headaches
• Growth of “things” can be overwhelming
• Observing the landscape of “things” can be a real challenge
• Visualization of “things” which are in a bad state and those which
may be trending towards a bad state is critical.
– Who cares about the stuff in a healthy state!?!
Example: Host List
Re-design…
New Concept:
List View
Combines:
• table visualization
• filter for defining a common
‘where’ clause
• discrete query by column
• dashboard linking for drill-
down
Deliver new “Host List View” as
an instantiation of this concept
List View
Configure Column Data
Configure Column Thresholds
Configure Column Links
List View: In Action
Cool!
Custom List Views with Drill-Down Links.
Did I mention I really don’t like tables??
Don’t tell me…
you want a pie
chart?
Landscapes of Data
Richer visualization…
Abstrac
t
Concrete
Time
That would be pretty awesome!
What happened to the sharing part?
Community
Simplify Sharing of Expertise
• Easier import/export of dashboards
• More enablement around creation and contribution of dashboards
– Tie into Telegraf plug-in creation
• Define/adopt specification for collections of assets
– Dashboard(s)
– Alert threshold, etc.
• Host community collaboration site for sharing
Dashboard Land!
Nirvana!
Thank You

Chronograf Present and Future

Editor's Notes

  • #10 Why? ”a large spike in metric of service A tells me there is a problem, and the log lines provide additional context"?
  • #17 another example outside of our “domain”  email inbox with 1000s of unread messages.
  • #18 Creating groupings of things allows you to view status at a higher level and drill into the problems faster. List of groups -> things in a group -> metrics about thing -> metrics about service related to thing
  • #19 Every "thing" has a different set of indicators of what could indicate an issue. Columns need to be able to be customized based on the things returned in the filter.
  • #29 Be sure to mention the creator of a service or app is the best person to know what needs to be monitored (can use influx as an example)