NOTE: We recommend Grafana for visualization of Prometheus metrics nowadays, as it has native Prometheus support and is widely adopted and powerful. There will be less focus on PromDash development in the future.


PromDash is a browser-based dashboard builder for Prometheus. It is a Rails application and stores its dashboard metadata in a configurable SQL backend.

PromDash screenshot

For installing and running PromDash, see the for more information. The instructions below assume that you have a running PromDash installation.

Registering a Prometheus server

Before you can graph data from a Prometheus server, you need to register it with PromDash. Under the "Servers" tab, click "New Server" and enter the name and URL of your Prometheus server. The URL needs to be in the form http://<host>:<port>/ - note the trailing slash. Click "Create Server". This server is now available for selection when configuring Prometheus graphs in a dashboard.

Creating a dashboard

To create a new dashboard, go to the "Dashboards" tab and click "New Dashboard". Enter a name and click "Create Dashboard". The dashboard name is now visible in the dashboards listing.

Configuring a dashboard

A PromDash dashboard consists of a number of widgets which are arranged in columns. Currently, two types of widgets are supported: Prometheus graphs and arbitrary iframes (including some special support for auto-adjusting URL parameters in Graphite graph links).

Global settings

The menu bar at the top of a dashboard lets you configure global options which have an effect on the dashboard as a whole:

  • Time range: Either use the + or - buttons to increase or decrease the global graph time range or type in a duration string of the form <number>[smhdwy].
  • Time window: Either use the << or >> buttons to shift the global graph time window backwards or forwards in time, or click on the input to select a specific date/time as the end of the graph range.
  • Manual refresh: The manual refresh button lets you trigger a refresh and redraw of all graphs on the dashboard.
  • Automatic refresh rate: This lets you set a periodic refresh interval.
  • Dashboard settings menu:
    • Columns: Choose the number of columns in which to display the dashboard's widgets.
    • Aspect ratio: For each widget, this sets an aspect ratio which will be automatically maintained across resizes.
    • Theme: Choose a light or dark look for you dashboard.
    • Color Scheme: Allows you to select the color scheme for lines in your graphs.
    • Resolution: A slider that selects the resolution with which to load data from Prometheus.
    • Show annotation tags: Select whether to show the UI for editing tags that specify which annotations to load and display in graphs. See the Annotations section.
    • Show template variables: Select whether to show the UI for editing global template variables for the dashboard. See the Using template variables section.
    • Encode entire dashboard in URL: Select whether to continuously update the browser URL as changes to the dashboard are made. The browser URL will then always reflect the current contents and state of the dashboard and may be sent around to exactly recreate the dashboard in the state when the link was generated.
  • Fullscreen mode: This displays the entire dashboard in fullscreen/TV mode, without any global menus. You may exit fullscreen mode by pressing "Escape".
  • Save Changes: This saves any changes to either the global dashboard settings or the settings of individual widgets. Without saving, any changes made to the dashboard are lost when reloading the page.

Adding Prometheus graphs

To add a Prometheus graph, click the "Add Graph" button on the dashboard.

To configure the graph, move the mouse over its widget area until a menu bar appears in the top left corner.


To configure which data to display, select the Datasources menu tab. It allows you to set one or multiple expressions to be graphed, along with the Prometheus server they should be queried from. The supported expressions are any standard Prometheus expression language expressions.

For each expression, you can select a graph axis to map the resulting data on, as well as a format string that specifies how the returned time series labels should be reflected in the legend. See later sections on how to add axes and format strings.

Time options

The time options let you configure a custom range and time window per graph. Note, however, that these are reset to the global time settings once the global ones are changed again.

Graph and axis settings

The graph and axis settings allow you to configure a title for your graph, as well as the preferred line interpolation and query resolution. Furthermore, you may add a second axis and configure various options for each axis (stacked vs. lines, axis scales, etc.).

Palette settings

The palette settings allow you to select a color scheme for drawing the lines and filled areas in graphs. If set, this overrides the global setting as long as the global setting remains unchanged.

Legend settings

The legend settings menu tab allows you to configure when to show the legend on a graph, as well as defining how it should be formatted.

By default, each time series will be displayed with their full name (metric name plus all label names and values) in the legend. If you have many time series or time series with many labels, this can quickly lead to unusably long legends.

The Legend format builder allows you to create a list of format strings which you can then assign to expressions in the Datasources tab. Each format string allows you to customize how results returned from an expression should be displayed in the legend. For each time series, the legend will show an interpolated version of the given format string. To reference specific label values in the format string, use double curly braces: {{label-name}}. For example: {{host}} - cluster {{cluster}}.

Format strings support filters. See the Filters section below for a list of currently available filters, expected inputs, and outputs.

The "Link to this graph" menu tab allows you to generate a link to a specific graph. This link will show the graph in a single-widget fullscreen view as it was configured when the link was generated. Subsequent changes to the underlying graph will not affect the linked version of the graph.

Adding inline frames

You can add arbitrary iframes as widgets to a PromDash dashboard by clicking the "Add Frame" button. You will be prompted to enter a URL to display, which may later be configured via the "Frame Source" menu tab. Similarly to Prometheus graph widgets, frame widgets allow you to configure a title and generate a link to the frame.

Automatic Graphite graph URL adjustment

PromDash has limited support for parsing and rewriting links to Graphite graph URLs. If you want to embed a Graphite graph via a PromDash frame, enable the option "this is a graphite graph" under the "Frame options" menu tab. PromDash will then automatically resize the Graphite graph to fit its widget by dynamically rewriting the frame URL. It will also add time controls which act similarly to those found in Prometheus graph widgets.

Using template variables

Template variables allow you to make generalized dashboards in which you set only certain variables each time you display them. For example, if you were interested in building a dashboard that shows statistics for a host, you would not want to build a separate dashboard for each host. Instead, PromDash lets you define the dashboard once and then create global template variables for the changing parts of your dashboard. In the case of the host dashboard, you might create a template variable called host and then let the user set this variable to the name of the host which they are currently interested in. The value of the variable value may then be inserted into relevant text input fields in a widget's settings.

A variable is interpolated into a text input field by enclosing it with a set of two curly braces: {{variable-name}}. The text inputs that support template variable interpolation are:

  • The widget title.
  • The expression input fields in a Prometheus graph.
  • The URL input of an iframe widget.
  • Annotation tag input fields.

In the example of a host dashboard, we could interpolate the host variable into each widget's Prometheus expressions, like:


...and set the title to:

CPU time for {{host}} in nanoseconds

Besides creating and setting variables via the user interface, you may set existing variables or even define new ones via the URL:


In the example of the host dashboard, the URL could look like this:!?

Template variables support filters. See the Filters section below for a list of currently available filters, expected inputs, and outputs.


Filters can be used in all places where variable interpolation is supported, e.g. in legend format strings or template variables. The format is {{variable | filter}} and the following filters are currently available:

  • toPercent: Input: 0.5; Output: 50%
  • toPercentile: Input: 0.5; Output: 50th
  • hostnameFqdn: Input:; Output:
  • hostname: Input:; Output: your-prometheus-endpoint
  • regex: If job == prometheus, {{job | regex:"pro":"faux"}} => fauxmetheus

Filters are chainable, so {{label | filter1 | filter2}} will apply filter1 to label, and then apply filter2 to that result.


PromDash allows you to load timestamped annotations from an external service and display them in graphs as vertical lines:

Annotation screenshot

Configuring annotations

This section assumes that you already have a working annotation server. For details on building one, see the Annotations API section below.

The URL that PromDash should query for annotations is set via the ANNOTATIONS_URL environment variable.

To enable annotations on a dashboard, first select "Display annotation tags" in the global dashboard settings. Each input can hold an arbitrary number of comma-separated tags that will be used to query your designated annotations server. Once you have added your tags you can hide the annotation list by un-checking annotations setting in the global dashboard settings.

Hovering over an annotation on a graph will display the annotation's message:

Annotation global config screenshot

Annotations API

Although we do not distribute a publicly available annotations server yet, it is easy to build one.

The URL from which PromDash fetches annotations has the following format:


The parameters have the following meaning:

  • ${ANNOTATIONS_URL}: this is the configured ${ANNOTATIONS_URL} environment variable.
  • until: This is the graph's end time as a Unix timestamp in seconds.
  • range: This is the graph's range (stretching back from the end time) in seconds.
  • tags[]: This is a list of the tags for which annotations should be loaded.

Each tag input on your dashboard represents a separate query. For example, if you have three tag inputs, three requests will be made to your annotations server.

The until and range parameters define the time window displayed by the graph. Annotations that are returned to PromDash that do not fall within this window will not be rendered.

The JSON payload returned from the annotations server needs to conform to this schema:

  posts: [
      created_at: 111232553, // UNIX timestamp
      message: "annotation message here"
    // ... more tags

All returned annotation tags will be rendered on every graph. There currently is no support for rendering specific tags on only a subset of graphs.

Embedding a dashboard

PromDash has support for embedding dashboards into other pages via iframes. To get an embeddable view for a PromDash dashboard, add embed to the URL of a dashboard as the first path element:


This view shows the dashboard's graphs in a carousel-view, one graph at a time. It omits the menu structure, the dashboard title, and global controls.

In the example of the host dashboard, the embeddable URL could look like this:!?

Pushing dashboard representations as JSON via HTTP

It's possible to retrieve and update the content of existing PromDash dashboards via the JSON format over HTTP. For example, this allows you to manage the authoritative version of your dashboard content in source control, while still using PromDash for the display.

Getting a dashboard's JSON representation

curl http://<promdash-url>/<dashboard-slug>.json > dashboard.json

Updating a dashboard's JSON representation

curl -H 'Content-Type: application/json' -X PUT -d @dashboard.json http://<promdash-url>/<dashboard-slug>.json

Specifying servers by URL

Any occurrences of serverID in the JSON format may now be optionally replaced by a serverURL property, specifying a server by its URL instead of by its internal database ID. A serverURL property needs to match the URL of an existing configured server in PromDash exactly - otherwise a 422 Unprocessable Entity HTTP error is returned.

Be aware that if you GET the JSON representation of a dashboard, it will always show serverID fields, and never serverURL, since any uploaded JSON containing serverURLs is transformed immediately to serverIDs before being saved server-side.


Add documention about:

  • fullscreen / fullscreenTitle URL parameters
  • document the dashboard JSON format