Grafana Associate key terms

A configured connection that tells Grafana where to fetch data, such as Prometheus, Loki, or a SQL database. Each one holds connection settings, authentication details, and query options used by every query built against it.

The panel-level interface where you write the request sent to a data source. Its fields change with the plugin, ranging from a PromQL expression box to structured SQL builders with table and column pickers.

Prometheus's query language for selecting and aggregating time series. Grafana users write it inside the Prometheus query editor, often combining selectors, rate functions, and aggregation operators like sum by to shape graphs.

A data source request that returns a single value per series at one point in time, rather than a range of samples. Useful for stat panels, tables, and gauges that show current state.

A request that returns many samples across a time window, producing the series drawn by time-based panels. The window comes from the dashboard time range, and sample spacing is controlled by the step or interval.

The spacing between returned data points in a range query, often set automatically from the dashboard time range and panel width. Too small a step overloads the data source; too large a step hides spikes.

A panel setting capping how many points are requested or rendered per series. Grafana uses it with the time range and panel width to compute an automatic interval, reducing load and overplotting.

A data source or query setting that sets a floor on the computed step, so queries never request samples more frequently than the configured value. It protects backends from excessive load on wide panels.

Collapsible per-query settings in the panel editor covering things like max data points, min interval, relative time override, cache timeout, and whether the query is hidden from the legend.

A server-side transformation written in Grafana's expression language that operates on other query results, enabling math, filtering, and reducing without changing the underlying data source query.

The access mode where the browser sends requests to the Grafana server, which forwards them to the data source. It keeps credentials private and avoids browser cross-origin restrictions.

An installable extension that adds data sources, panels, or apps to Grafana. Data source plugins define how to query a backend, while panel plugins define new visualization types.

A saved collection of panels arranged on a grid, sharing a time range, variables, and refresh behavior. Dashboards are the primary artifact users build, browse, and share in Grafana.

A single visualization block on a dashboard, combining one or more queries with display options like thresholds, units, and legends. Panels can be resized, moved, duplicated, and inspected.

The rendering chosen for a panel, such as time series, stat, gauge, bar chart, table, or heatmap. The right choice depends on whether data is over time, a single value, or categorical.

The default panel for plotting numeric values against time as lines, bars, or points. It supports multiple series, legend controls, thresholds, and axis configuration.

A panel that displays one prominent value per series, optionally with sparkline, color mode, and background coloring driven by thresholds. It suits current totals, counts, and health indicators.

A panel showing a value as a filled arc against a defined range, with color segments from thresholds. Good for capacity and utilization metrics where proximity to a limit matters.

A panel that renders query results as rows and columns, supporting column formatting, cell coloring, sorting, and field overrides. It suits logs, inventories, and multi-field query output.

A panel that plots value distribution over time as colored buckets, commonly used for latency histograms from Prometheus. Color intensity shows how many observations fell in each bucket.

Value boundaries that change a panel's color or trigger alert states. In stat, gauge, and time series panels they paint values green, yellow, or red according to configured steps.

The measurement label applied to displayed values, such as bytes, seconds, percent, or requests per second. Grafana formats and scales numbers automatically once a unit is chosen.

The panel area listing series names with color swatches, and often last, min, max, or mean values. It can be placed at the bottom, right, or hidden, and supports click-to-isolate behavior.

Per-field display settings in the panel editor controlling unit, decimals, min and max, color scheme, and thresholds. Overrides apply them to specific fields or series matching a rule.

A rule in the panel editor that changes field options for matching fields, series, or names, leaving other fields untouched. Overrides target by name, regex, query, or value-based conditions.

A processing step applied to query results before visualization, such as joining, filtering, renaming, or calculating a new field. Multiple transformations chain in order in the panel editor.

A dashboard-level placeholder whose selected value is substituted into queries, panel titles, and links. Variables let one dashboard serve many services, hosts, or environments without duplication.

A variable whose options are populated by running a query against a data source, such as listing label values or table columns. It can refresh on dashboard load or on time range change.

A variable whose allowed values are typed manually as a comma-separated list, optionally with display labels. Useful for fixed choices like environment names or predefined tiers.

A variable setting that lets users select several options at once. Queries must use the correct formatting syntax so the values expand into a valid list for the target data source.

A variable whose option query depends on another variable's current selection, narrowing choices as the parent changes. Common for region then host, or cluster then namespace hierarchies.

The window of time a dashboard or panel displays, set absolutely with dates or relatively with expressions like now-6h. It is passed to data source queries and can be overridden per panel.

The dashboard control for choosing the time range and refresh interval, including relative shortcuts and custom absolute ranges. Changing it re-runs queries for all panels that follow it.

How often a dashboard automatically re-runs its queries, chosen from the time picker such as every 30 seconds. It affects load on data sources and is often disabled for slow queries.

A Grafana view for ad hoc querying without building a dashboard, splitting queries side by side and showing raw results. It is the usual first stop for debugging a data source or writing a query.

The serialized model of a dashboard, including panels, queries, variables, and layout. It can be exported, imported, version-controlled, and provisioned as code.

Loading dashboards, data sources, and alert rules from configuration files at startup rather than through the UI. Provisioned items are typically read-only in the interface and managed in version control.

A container that groups dashboards and controls access through permissions. Folders organize large instances and determine who can view or edit the dashboards inside them.

A saved, static copy of a dashboard's rendered data that can be shared by link without access to the original data sources. Useful for incident reports and external sharing.

A URL that opens a dashboard, optionally with the current time range and variable selections embedded. Links can be shortened or made absolute for use outside the Grafana instance.

A permission level assigned to users or teams, such as Viewer, Editor, or Admin, determining what they can see and change. Roles combine with folder permissions to gate dashboard access.

A tenant inside a Grafana instance that isolates dashboards, data sources, users, and settings from other tenants. A single user can belong to several organizations with different roles.

A non-human identity with a token used for automation, API calls, and provisioning. It holds roles like a user but has no interactive login, making it suitable for CI pipelines.

A timestamped event marker overlaid on dashboard graphs, such as a deployment or incident. Annotations can be created manually, via API, or queried from a data source.