Change Details

=== Background https://phabricator.wikimedia.org/T422943#11998988 indicates we will not have Custom Hooks available for GrowthBook experiments, thereby necessitating backend validation during ingestion on the TK backend. [[ https://docs.google.com/spreadsheets/d/1luvl6TPX_rgYdS8C_912FoCvyFqhQSbQnquP-xfuUG0/edit?gid=31143002#gid=31143002 | Canonical GrowthBook customization/validation spreadsheet ]] captures validation rules. === Description Per [[ https://docs.google.com/spreadsheets/d/1luvl6TPX_rgYdS8C_912FoCvyFqhQSbQnquP-xfuUG0/edit?gid=31143002#gid=31143002 | canonical spreadsheet tracking validation rules ]], add the following backend validation rules: [] Tracking Key (machine readable name) - must be unique + varnish pattern - GB experiment prevails when two experiments (GB and TKUI) have same machine-readable name [] Related OKRs (defined as tag in GB) - Check if present via local API for OKRs - We could use the local API we have for OKRs to check if at least one OKR if present as a tag [] Teams (defined as tag in GB) - Check if present via local API for Teams - Same thing as for OKRs. In this case we would need to build that local API + database table for teams [] Stream Name - required custom field validated via Action API [] Traffic per Wiki - check if present - check here that 0 <= length(wikis) <= 100 [] Risk Level - check via current validation logic - "https://wikitech.wikimedia.org/wiki/Test_Kitchen/Contextual_attributes#Privacy_considerations: If risk_level === 'Risk assessment pending' => experiment cannot run but there is no way to enforce that in GB Here we will need backend validation anyway" [] Security and Legal Review - check via current validation logic - Mandatory is risk_level is Tier 2: Medium or Tier 1: High [] When any of the validation rules above fails, the experiment status is set to `Draft` as a way to "stop" it === Acceptance Criteria [] Validation rules are implemented [] Relevant tests added/updated

=== Background https://phabricator.wikimedia.org/T422943#11998988 indicates we will not have Custom Hooks available for GrowthBook experiments, thereby necessitating backend validation during ingestion on the TK backend. [[ https://docs.google.com/spreadsheets/d/1luvl6TPX_rgYdS8C_912FoCvyFqhQSbQnquP-xfuUG0/edit?gid=31143002#gid=31143002 | Canonical GrowthBook customization/validation spreadsheet ]] captures validation rules. === Description Per [[ https://docs.google.com/spreadsheets/d/1luvl6TPX_rgYdS8C_912FoCvyFqhQSbQnquP-xfuUG0/edit?gid=31143002#gid=31143002 | canonical spreadsheet tracking validation rules ]], add the following backend validation rules: [] Tracking Key (machine readable name) - must be unique + varnish pattern - GB experiment prevails when two experiments (GB and TKUI) have same machine-readable name [] Related OKRs (defined as tag in GB) - Check if present via local API for OKRs - We could use the local API we have for OKRs to check if at least one OKR if present as a tag [] Teams (defined as tag in GB) - Check if present via local API for Teams - Same thing as for OKRs. In this case we would need to build that local API + database table for teams [] Stream Name - required custom field validated via Action API [] Traffic per Wiki - check if present - check here that 0 <= length(wikis) <= 100 [] Risk Level - check via current validation logic - "https://wikitech.wikimedia.org/wiki/Test_Kitchen/Contextual_attributes#Privacy_considerations: If risk_level === 'Risk assessment pending' => experiment cannot run but there is no way to enforce that in GB Here we will need backend validation anyway" [] Security and Legal Review - check via current validation logic - Mandatory is risk_level is Tier 2: Medium or Tier 1: High === Acceptance Criteria [] Validation rules are implemented [] When any of the validation rules above fails, the experiment status is set to `Draft` as a way to "stop" it [] Relevant tests added/updated

=== Background https://phabricator.wikimedia.org/T422943#11998988 indicates we will not have Custom Hooks available for GrowthBook experiments, thereby necessitating backend validation during ingestion on the TK backend. [[ https://docs.google.com/spreadsheets/d/1luvl6TPX_rgYdS8C_912FoCvyFqhQSbQnquP-xfuUG0/edit?gid=31143002#gid=31143002 | Canonical GrowthBook customization/validation spreadsheet ]] captures validation rules. === Description Per [[ https://docs.google.com/spreadsheets/d/1luvl6TPX_rgYdS8C_912FoCvyFqhQSbQnquP-xfuUG0/edit?gid=31143002#gid=31143002 | canonical spreadsheet tracking validation rules ]], add the following backend validation rules: [] Tracking Key (machine readable name) - must be unique + varnish pattern - GB experiment prevails when two experiments (GB and TKUI) have same machine-readable name [] Related OKRs (defined as tag in GB) - Check if present via local API for OKRs - We could use the local API we have for OKRs to check if at least one OKR if present as a tag [] Teams (defined as tag in GB) - Check if present via local API for Teams - Same thing as for OKRs. In this case we would need to build that local API + database table for teams [] Stream Name - required custom field validated via Action API [] Traffic per Wiki - check if present - check here that 0 <= length(wikis) <= 100 [] Risk Level - check via current validation logic - "https://wikitech.wikimedia.org/wiki/Test_Kitchen/Contextual_attributes#Privacy_considerations: If risk_level === 'Risk assessment pending' => experiment cannot run but there is no way to enforce that in GB Here we will need backend validation anyway" [] Security and Legal Review - check via current validation logic - Mandatory is risk_level is Tier 2: Medium or Tier 1: High [] When any of the validation rules above fails, the experiment status is set to `Draft` as a way to "stop" it === Acceptance Criteria [] Validation rules are implemented [] When any of the validation rules above fails, the experiment status is set to `Draft` as a way to "stop" it [] Relevant tests added/updated