Building dashboards
A dashboard walkthrough builds a complete board around one operational question, tile by tile. Each article in this section is named for the question its dashboard answers, so start with the one that matches a question you already have rather than working through them in order.
Every tile is documented on its own page, so a walkthrough can be followed end to end or raided for the two or three tiles you actually need.
Dashboard walkthroughs
Each walkthrough builds one complete board, naming the data sources it needs and linking to the recipe behind every tile.
- Deployment health and risk: Deployment success rate, change failure rate, unresolved critical issues and a composite risk score.
- Major incidents: Active P1s, ticket and alert momentum, affected services and assignment group load.
- On-call alert noise: Alert volume, out-of-hours distribution, actionability, MTTA and MTTR.
- DORA metrics: Deployment frequency, lead time, change failure rate and MTTR.
What a walkthrough covers
- The data sources it draws on: Named up front, along with what each one contributes. Most walkthroughs combine two, because the interesting questions sit across a system of record and a telemetry source.
- A tile at a time: Each tile gets a short description of the signal it carries and a link to the recipe that builds it.
- The order to build in: Tiles are independent, with one exception. Composite tiles that read from published KPIs need the tiles feeding them built first, and the walkthrough says which one to leave until last.
Before you start
You need at least one data source connected. See adding a data source if you have not set one up yet.
If you have not built a dashboard before, work through building a dashboard first. The walkthroughs assume you can already add a tile and open the tile editor.
Get more out of the finished dashboard
A walkthrough gets you the tiles. Three things turn that into something a team uses daily.
- Set the timeframe and scope deliberately: The dashboard timeframe drives every tile that accepts one, so the same board can answer a live question and a quarterly one.
- Add monitors to the tiles that carry a threshold: A monitor turns a value into a state, so the dashboard tells you something is wrong without anyone reading it.
- Publish the values worth reusing as KPIs: A published KPI becomes a data stream other dashboards can select, which is how composite tiles get their inputs.
Walkthroughs and prebuilt dashboards
Several plugins ship prebuilt dashboards covering the same ground, and the plugin article says whether yours does. Those are the fastest way to see your data and are worth turning on first.
Build your own when you need signals from more than one tool on the same board, when you want to change how a metric is derived, or when you want to monitor a value and publish it for other dashboards to reuse.