StackState Self-hosted v5.0.x

This page describes StackState version 5.0.

Go to the documentation for the latest StackState release.


The analytics environment allows you to directly query the 4T data model. The analytics environment uses the StackState Scripting Language (STSL) as the basis for querying StackState, so you can build and test your StackState scripts here.

Queries created in the analytics environment can be used to investigate issues, automate processes and build reports. Here are some examples of queries that you could execute:

  • Get the names of all pods running in a namespace.

  • Determine the maximum latency of a service since yesterday.

  • Find all machines indirectly connected to a set of APIs.

  • Show which databases have been updated since last week.

The analytics environment

If your StackState user has the permission access-analytics, then you can access the analytics environment from the main menu. This will be available by default for all power users and admins. There are also places in the user interface where you can be directed to analytical environment if some data is available in the form of a query, a link will then say "Open in Analytics".

The analytics environment is divided into two sections:

  • The query you want to execute on the left

  • The results and query history on the right.

When executing a query for the first time, the result of the query is displayed in preview form if a preview is available for the type of data that is requested, for example a metric chart or a topology view. If no preview is available, the data will be shown in JSON form.

Every query that you have executed in StackState is shown in the query history, together with the query result at that point in time.


Results of queries are typically displayed in raw JSON form, unless there is a preview available. Previews are currently available for:


In the analytics environment, you use a combination of the StackState Scripting Language (STSL) and the StackState Query Language (STQL) to build your queries and scripts.

A query is a regular STSL script. For example, when you run the query: 1+1 you will get the result 2.

As a part of an STSL script, you can invoke the StackStake Query Language (STQL). A simple example of an analytical query that uses both STSL and STQL is:

Topology.query('environment in ("Production")').components()

Topology.query is a regular script function that takes ab STQL query (environment in ("Production") in the above example) as an argument. The .components() at the end, is a so-called builder method. This ensures that only the components, and not the relations between these components, are retrieved from the topology.

The combination of STSL and STQL allows you to chain together multiple queries. The following example gets all metrics of all databases in the production environment for the last day:

    .query('environment in ("Production") AND type = "Database"')
    .thenCollect { metricStream -> 
            .aggregation("95th percentile"", "15m")

This analytical query first gets all metrics streams of components from the Production environment that are of the type Database. The result of that query is then used to build up telemetry queries against these metric streams.

The full list of available functions can be found in the Script API documentation. To learn about chaining, see async script results.

Example queries

Below are some queries to get you started with an example of their expected output. You can find more examples in the StackState UI Analytics environment itself.

Find the number of relations between two components

Topology.query('name IN ("Alice", "Bob")')

Compare the Staging environment to the Production environment

Topology.query('environment = "Staging"')
    .diff(Topology.query('environment = "Production"'))
    .then { diff ->
            .collect { comp -> }

Predict disk space of a server for the next week

Prediction.predictMetrics("linear", "7d",
    Telemetry.query("StackState metrics", 'host="lnx01" AND name="diskspace" AND mount="/dev/disk1s1"')
        .aggregation("min", "1d")
        .start("-4w") // based on last month
).predictionPoints(7).then { result -> resut.prediction  }

See also

Last updated