Common workflows

Learn the standard tool sequences for running tests, creating monitors, querying metrics, and investigating incidents on the platform.

These recipes show the tool order for the tasks agents run most often. Each step names the tool to call.

Before a multi-step task, call platform_guide with the relevant topic (for example test_creation or metrics). It returns the up-to-date sequence and prerequisites straight from the server.

Run a test

1
Find the test

list_tests with a product slug, or search_collections to find it in the folder tree.

2
Check what it needs

get_test_detail to see required input variables and owners.

3
Build and run a campaign

build_campaign to assemble the run (pass a variants array to pick an environment), then run_campaign.

4
Read the result

poll_execution until it finishes, then get_execution_detail for per-test pass/fail data.

Create a monitor

1
Pick the test and environment

list_tests for the test id, and list_variable_groups for the variable group that holds your environment.

2
Create the monitor

create_monitor with the product, test id, variable group, and schedule.

3
Activate or pause

toggle_monitors to turn monitors on or off in bulk.

Query metrics

1
Explore the schema

query_metrics with SHOW MEASUREMENTS and SHOW FIELD KEYS to discover what is collected.

2
Query the series

query_metrics with a SELECT to read KPI time series, then compute_slo or correlate_metrics for deeper analysis.

Investigate incidents

1
Check overall health

get_incident_counter for a quick critical / warning / closed count.

2
List and inspect

list_incidents filtered by status, then get_incident for detail.

3
Resolve

close_incident once the underlying issue is handled.

What's next?

Tools reference

Look up the exact parameters for each tool above.

Core concepts

Variables, variants, edit budgets, and the safety model.

Last updated on