Bug Bash 10.1 - Metrics Squad

[itsmylife]

Test Environment Information
Release: 10.1
Test Environment: https://testymctestface.grafana-dev.net
Browser and browser version:

Mandatory for NEW features or enhancements

• Transformations: New design
  • Transformation redesign #70834
• Prometheus browser resource cache
  • Enabled by default in 10.1.x
  • Duplicate requests display as cached in network inspector, within the time frame defined by the cache setting in the config.
  • See doc for more info
  • See an example
• Prometheus: Template variables query editor UI update under "label values" query type
  • Prometheus: Variable query editor improvements #69884
  • Use the metric select in the Prom query variable editor.
  • Allow the metric to be filtered with optional label filters.
  • Allow the metric to be clearable.
  • When a metric is selected, filter the options of required label name and select dropdown
  • Add variables to the options in the select dropdowns (labels and metric for the label_values query type).
• Dataplane: Azure monitor metrics time-series data moved to dataplane
  • Has feature flag enabled by default, would expect Datasource squad to be checking for impact
  • Azure: AzureMonitorMetrics - change response to be dataplane compliant #69308
• Prometheus: Config overhaul part two, auth and DataSourceHttpSettings overhaul
  • Prometheus: Config overhaul part two, auth and DataSourceHttpSettings overhaul  #71250
  • Feature toggle turned on in dev and staging. Will be ready to test by Friday, Aug 4th 2023
  • The Prometheus configuration page is using the new grafana/experimental Auth component
  • This means the DataSourceHTTPSettings component has been refactored so the following inputs have been added or moved.
    • description component
    • connection component
    • advanced settings component
    • add sigV4 auth to custom auth
    • add Azure auth to custom auth

Hidden behind feature toggle that defaults to false

• InfluxDB backend migration validation InfluxDB: Backend migration validation #65045

Test scenarios:

Transformations: PASSED

• Create a new dashboard
• Create a new visualization
• Select any datasource
• Click the “Transform” tab
• You can now see the new transformations picker
• You can click on the categories, just below the search input. This will filter the list of transformations

Prometheus browser resource cache: PASSED

• Command + k -> go to data sources
• Find a prom data source
• Check that the cache level dropdown is visible, and populated (will be “low” by default)
• Change HTTP method to GET
• If you want to wait around a bit, check the different cache values, see the doc for target assertions
• Note that metadata has different cache durations then other endpoints
• Open a panel in explore or dashboard
• Write a promQL query via querybuilder or code editor
• Whenever a label/…/values, /labels, or metadata (or series) call is made with the same non-time parameters in the same cache interval (defined by cache level in config), it should display in network inspector as fetched from disk cache
• ** Note, make sure your browser cache is not disabled 🤦

Prometheus template query editor variable label values option PASSED

• To start:
  • Use “Prometheus Demo Robust Perception” data source or any other Prometheus data source
  • Create dashboard using “Prometheus Demo Robust Perception” data source
  • Go to dashboard settings > Variables
  • Select ‘add variable’
  • Scroll to Query options and select “Prometheus Demo Robust Perception” data source
  • All tests will be performed with query options.
• Scenario 1. Filter label names by metric
  • Make sure you have selected a Prometheus data source.
  • In Query options, choose query type “label values.”
  • View the label options in the Label dropdown.
  • Choose a metric in the metric dropdown.
  • See that the labels have been filtered by the selected metric.
• Scenario 2. Filter metric by label filters
  • Make sure you have selected a Prometheus data source.
  • In Query options, choose query type “label values.”
  • View the options for the metric select below the label names input.
  • Choose label filters by selecting a label name and label value.
  • See that the metric options have been filtered by label filters.
• Scenario 3. Use the variable in metric select
  • Make sure you have selected a Prometheus data source for all variables you create in the following steps.
  • Create a new variable named ‘metrics’.
  • Choose the Metrics query type.
  • Enter the following metric regex: go
  • See that there are variable options at the bottom of the screen.
  • Save the variable.
  • Go to your dashboard and select a value from the dropdown for metrics.
  • Create a new variable.
  • Select the label values query option.
  • Select your variable in the metric dropdown
  • See that the label options have been filtered by the metric variable.
• Scenario 4. Use variable in label filter
  • Create a new variable named ‘.
  • Choose the Label names query type.
  • Select a label name.
  • See that there are variable options at the bottom of the screen.
  • Save the variable.
  • Go to your dashboard and select a value from the dropdown for label names.
  • Create a new variable.
  • Select the label values query option.
  • In Label filters, select the variable in the dropdown with the ‘Select label’ placeholder.
  • See that the label values in the Label filters have been filtered by the label name variable.

Prometheus configuration overhaul

• Scenario 1. View a saved/configured data source
  • View a data source that has been saved
  • Home > Connections > Data sources >
  • Check the auth component
    • Is the correct type of auth selected?
    • Are the input values correct?
  • Do you have custom HTTP headers?
    • If so, are they correctly listed below the auth component?
  • Check the Advanced HTTP settings component.
  • Did you add either of the following inputs?
    • Allowed cookies
    • Timeout
    • Are they listed correctly?
  • Check the connection string.
    • Is the connection string to the Prom data source correct?
  • Can you query the data source in the Explore page?
  • Special review (these are not turned on in dev for the bug bash but you may test them locally if you so wish by turning on the appropriate flags. They will be additionally tested in ops.)
    • The following auth types can be selected when the feature toggles are turned on for them.
    • SigV4 auth
      • Known issue: the inputs are too wide
    • Azure auth
• Scenario 2. Configure a new Prometheus data source.
  • Configure a new data source with the following:
  • Connections string and some form of authentication.
  • Save and query the data source in Explore.
  • Return to the data source configuration page and see that all the inputs are correct.

Issues found:

• Transformations: “Format Time” transformation has no image
  • PR: Transformations: Add illustration for formatTime #72673
  • The PR is already merged and it needs to be ended up in the testing environment.

Go/No-Go (if applicable): link to Go/No-Go doc

[grafanabot]

@itsmylife please add one or more appropriate labels. Here are some tips:

• if you are making an issue, TODO, or reminder for yourself or your team, please add one label that best describes the product or feature area. Please also add the issue to your project board. 🚀

• if you are making an issue for any other reason (docs typo, you found a bug, etc), please add at least one label that best describes the product or feature that you are discussing (e.g. area/alerting, datasource/loki, type/docs, type/bug, etc). Our issue triage doc also provides additional guidance on labeling. 🚀

Thank you! ❤️