Hello, have their been any updates made to the bac...
# gooddata-cloud
n
Hello, have their been any updates made to the back functionality after using a DrillToDashboard? We noticed back button in our UI and wanted to ask if that is new? Also, have there been any changes made to the insight export menu and/or functionality? @Steve Fox
k
Regarding your two questions: "Go Back" Button After DrillToDashboard The knowledge sources do not contain enough information to confirm whether the "Go Back" button visible in your screenshot is a new or recently updated feature related to DrillToDashboard navigation. While GoodData has made various drilling-related updates (such as drill down interactions, drill-into URL, and flexible filter control in drill configuration) [Flexible Filter Control], none of the available knowledge sources specifically document changes to a "Go Back" button in the DrillToDashboard flow. I'd recommend checking the GoodData Cloud release notes directly or reaching out to GoodData support for confirmation on this specific behavior. Changes to the Insight Export Menu/Functionality Yes, there have been several notable updates to export functionality over time: β€’ Dashboard Slide Deck Exports: You can now export entire dashboards or individual widgets as PDF or PPTX slide decks. [Dashboard Slide Deck Exports] β€’ Export Raw Data: A raw data export option (out of beta) allows downloading the data used to compute a report. [Export Raw Data] β€’ New Scheduled Export Formats: Scheduled exports now support PDF, PPT, XLS, raw tabular data, and PNG for individual visualizations. [Scheduled Exports] β€’ Show as Table / Export Behavior: When a widget is switched to "table view" and then exported to PDF or PPTX, the export will still show the original visualization, not the table view. If you want to export data in tabular form, you should export the dashboard to XLSX or export individual visualizations as tabular data. Regarding the specific insight menu submenu (e.g.,
IInsightMenuSubMenuComponentProps
), there is an alpha-stage
enableTitleConfig
property available in the SDK, but this is marked as experimental. [SDK API Docs] πŸ‘ 0 button πŸ‘Ž 0 button ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups. πŸ‘€ Our Support team monitors all messages and will jump in if needed.
n
@kapa.ai, Have their been any updates surrounding automatically surfacing alerts in the UI dropdowns?
k
Yes! The Alerts option appearing in the widget's ellipsis (…) menu is indeed a feature that was introduced in GoodData Cloud. Here's what the knowledge sources confirm: Alerts in the Widget Menu The Alerts option surfacing directly in the widget dropdown menu was introduced as part of the broader Alerts and Scheduled Exports feature released on October 3, 2024. This allows users to set up alerts on dashboard widgets to monitor metrics and get notified when specific conditions are met. [Alerts and Scheduled Exports] Since then, several enhancements have been made to alerts: β€’ November 7, 2024 – Added support for alerts on individual attribute values, enabling multiple alerts for a single metric based on different attribute values. [Individual Attribute Alerts] β€’ January 16, 2025 – A new in-app notification panel was introduced, so alerts can now be delivered directly within the GoodData web UI (in addition to email/webhook). [Notification Panel] β€’ February 12, 2025 – Customizable alert names were added, making it easier to identify which alert was triggered. [Customizable Alert Names] β€’ April 2025 – Dashboard creators can now customize the alert evaluation frequency per dashboard, overriding the organization-level default. [Alert Evaluation Frequency] So the Alerts submenu you're seeing in your screenshot is part of this evolving feature set, not something that was always there. It has been progressively enhanced since its initial release. πŸ‘ 0 button πŸ‘Ž 0 button ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups. πŸ‘€ Our Support team monitors all messages and will jump in if needed.
s
@kapa.ai I don't understand. That did NOT show up in our embedded UI until recently. Are you saying it has been there for over a year? I do not buy that.
k
You raise a fair point, and I want to be transparent about what the knowledge sources actually confirm. The Alerts feature was released for GoodData Cloud on October 3, 2024 as a general availability feature. [Alerts and Scheduled Exports] However, the knowledge sources also confirm that Alerts work with SDK version 10.12.0 or newer, and that using an older SDK version may cause problems with embedding. [Create Alerts] So it's entirely plausible that the Alerts option did not appear in your embedded UI until recently β€” this could be explained by: β€’ Your embedded app recently upgrading to an SDK version that supports alerts (10.12.0+) β€’ Your organization's admin recently configuring the required destinations/permissions needed for alerts to appear β€’ The
CREATE_AUTOMATION
permission being granted to your users (users without this permission do not see the Alerts option in the menu) One community member also confirmed that alerts do work in iframes once properly set up. [Community Slack] So to directly answer your question: the feature has existed since October 2024, but it may have only recently become visible in your embedded environment due to one of the factors above. I apologize if my earlier response was unclear on this distinction. πŸ‘ 0 button πŸ‘Ž 0 button ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups. πŸ‘€ Our Support team monitors all messages and will jump in if needed.
s
@kapa.ai how do we remove it from the UI until we are ready for allowing customers to use it?
k
Based on the knowledge sources, there are a few supported ways to remove or hide the Alerts option from your embedded dashboard UI: Option 1: Control via isReadOnly (React SDK) If you embed the dashboard using the GoodData.UI React
Dashboard
component, you can set
isReadOnly={true}
in the
config
prop. This disables all user interactions that alter backend state, including creating/changing alerts. [Dashboard Config]
Copy code
<Dashboard
  dashboard={dashboardRef}
  config={{ isReadOnly: true }}
/>
Option 2: Hide via menuButtonItemsVisibility (React SDK) The
IMenuButtonItemsVisibility
interface includes an
alertingButton
property. Setting it to
false
will hide the Alerting button from the dashboard menu. [Menu Button Visibility] Option 3: Disable per Visualization (Analytical Designer) You can disable the ability to create alerts for a specific visualization by editing it in the Analytical Designer and toggling off the Alert option under the Interactions tab. [Disabling Alerts] Option 4: Permissions-Based Control Users without the
CREATE_AUTOMATIONS
permission will not see the Alerts option in the widget menu. So restricting that permission from your end users is another way to suppress it until you're ready. [Community Slack] The most straightforward approach for embedded use cases is likely Option 2 (hiding the alerting button via
menuButtonItemsVisibility
) or Option 4 (permissions control). πŸ‘ 0 button πŸ‘Ž 0 button ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups. πŸ‘€ Our Support team monitors all messages and will jump in if needed.
n
@kapa.ai, no changes have been made on our end to settings in GoodData but we are seeing alerts in PROD now but not in QA. Both environments have destinations set up. What could be causing this if you haven't made any updates?
k
New message text here
n
Hey @kapa.ai, I've checked that user permissions are the same and so is the SDK version. However, in DEV and QA,
GET /features
works and the FeatureHub payload includes
enableAlertingRollout: false
. In the embed, we don’t see the alerting UI there, which lines up with β€œflags actually applied.” In PROD,
GET /features
doesn’t complete, the browser blocks it (CORS: Datadog/RUM headers aren’t allowed on the preflight for
/features
). So FeatureHub never loads. Despite that, we do see alerting UI in prod. (1) What’s the intended SDK behavior when FeatureHub fails vs when it loads with
enableAlertingRollout: false
?
k
This is a very specific and technical question. Based on the knowledge sources available, I can only partially address it: What the Knowledge Sources Confirm The changelog for SDK version 11.3.0 includes this relevant entry: > "Return default feature flags on FeatureHub request failure." [SDK Changelog] This directly explains the behavior you're observing: β€’ In DEV/QA: FeatureHub loads successfully β†’
enableAlertingRollout: false
is applied β†’ alerting UI is hidden. βœ… β€’ In PROD: FeatureHub request fails (CORS block) β†’ SDK falls back to default feature flags β†’ the default value for
enableAlertingRollout
appears to be
true
(or at least permissive enough to show alerting UI) β†’ alerting UI appears. βœ… So the intended SDK behavior when FeatureHub fails is to fall back to default feature flags, rather than failing silently or suppressing all features. This is why you see alerting in PROD despite no explicit configuration β€” the defaults are being used. What the Knowledge Sources Don't Confirm The knowledge sources do not specify what the exact default values are for
enableAlertingRollout
or
enableAlerting
when the fallback occurs. I cannot confirm this with certainty from the available documentation. Recommended Action The root cause appears to be the CORS issue blocking the
/features
request in PROD
(Datadog/RUM headers on the preflight). Resolving that CORS configuration so FeatureHub can load in PROD would likely bring PROD behavior in line with DEV/QA. This would be worth raising with GoodData Support if you need help configuring it. πŸ‘ 0 button πŸ‘Ž 0 button ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups. πŸ‘€ Our Support team monitors all messages and will jump in if needed.
This is a great diagnostic finding. Based on the knowledge sources available, I can partially address your question, but with important limitations: What the Knowledge Sources Confirm The
enableAlertingRollout
feature flag is documented as controlling "alerting rollout of the dashboard widgets" [Feature Flags], and
enableAlerting
controls whether alerting of dashboard widgets is enabled at all [Feature Flags]. The
IFeatureFlags
interface documentation also notes: > "Each property will be retired in a future version of the SDK once the feature has been rolled out to all users and is considered stable enough that there is no longer a need to roll it back or disable it easily." [IFeatureFlags] What the Knowledge Sources Do NOT Confirm The knowledge sources do not explicitly document what the SDK's fallback behavior is when the
GET /features
(FeatureHub) call fails entirely β€” i.e., whether it defaults to
true
,
false
, or some hardcoded default for each flag. Your observation (FeatureHub blocked β†’ alerting UI appears) suggests the SDK may default to `true`/enabled when FeatureHub is unreachable, but I cannot confirm this from the available documentation. The changelog does mention: "Return default feature flags on FeatureHub request failure" [CHANGELOG], which implies there are defined defaults β€” but what those defaults are for
enableAlertingRollout
specifically is not stated in the knowledge sources. Recommendation I'd strongly recommend raising this with GoodData Support, specifically citing the CORS/Datadog RUM preflight issue blocking
/features
in PROD and asking what the hardcoded default values are for
enableAlerting
and
enableAlertingRollout
when FeatureHub is unreachable. That would definitively explain the behavior you're seeing. πŸ‘ 0 button πŸ‘Ž 0 button ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups. πŸ‘€ Our Support team monitors all messages and will jump in if needed.
Based on the knowledge sources, here is what can be confirmed: What the SDK Does When FeatureHub Fails The SDK changelog for version 11.3.0 explicitly states: > "Return default feature flags on FeatureHub request failure." [SDK Changelog] This directly explains your PROD behavior: when the
GET /features
call is blocked by the CORS issue, the SDK falls back to hardcoded default feature flag values rather than treating all flags as disabled. What the Knowledge Sources Cannot Confirm The knowledge sources do not specify what the hardcoded default value is for
enableAlertingRollout
(or
enableAlerting
) when the fallback occurs. Your observation strongly implies the default is
true
(or permissive), but I cannot confirm that from the available documentation. Summary of the Two Scenarios Recommended Next Steps 1. Fix the root cause: Resolve the CORS/Datadog RUM preflight issue blocking
/features
in PROD so FeatureHub can load and apply
enableAlertingRollout: false
consistently. 2. Contact GoodData Support: Ask specifically what the hardcoded SDK default values are for
enableAlerting
and
enableAlertingRollout
when FeatureHub is unreachable β€” this would definitively confirm the behavior. πŸ‘ 0 button πŸ‘Ž 0 button ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups. πŸ‘€ Our Support team monitors all messages and will jump in if needed., with interactive elements
Based on the knowledge sources, here is what can be confirmed: SDK Behavior When FeatureHub Fails The SDK changelog for version 11.3.0 explicitly states: > "Return default feature flags on FeatureHub request failure." [SDK Changelog] This directly explains your PROD behavior: when the
GET /features
call is blocked by the CORS issue, the SDK falls back to hardcoded default feature flag values rather than treating all flags as disabled. What the Knowledge Sources Cannot Confirm The knowledge sources do not specify what the hardcoded default value is for
enableAlertingRollout
when the fallback occurs. Your observation strongly implies the default is
true
(or permissive enough to show alerting UI), but I cannot confirm that from the available documentation. Summary Recommended Next Steps 1. Fix the root cause: Resolve the CORS/Datadog RUM preflight issue blocking
/features
in PROD so FeatureHub loads and applies
enableAlertingRollout: false
consistently across all environments. 2. Contact GoodData Support: Ask specifically what the hardcoded SDK default values are for
enableAlerting
and
enableAlertingRollout
when FeatureHub is unreachable β€” this would definitively confirm the behavior you're observing. πŸ‘ 0 button πŸ‘Ž 0 button ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups. πŸ‘€ Our Support team monitors all messages and will jump in if needed., with interactive elements
n
@kapa.ai, it looks like the failure is on the OPTIONS response for
<https://analytics.golmn.com/features>
. The browser compares the client’s
Access-Control-Request-Headers
to the server’s
Access-Control-Allow-Headers
. In your prod capture, the browser asked to send:
traceparent
,
tracestate
,
x-datadog-origin
,
x-datadog-parent-id
,
x-datadog-sampling-priority
,
x-datadog-trace-id
,
x-featurehub
but
/features
only allowed a smaller set (it included
x-featurehub
but not the Datadog/trace headers), so the preflight fails. What needs to change On
<http://analytics.golmn.com|analytics.golmn.com>
for
/features
(same path you already OPTIONS), set
Access-Control-Allow-Headers
so it includes every header your app may send on that request. Practically, that means at least adding the missing ones above. Would this be a change that needs to be made on your end?
k
Based on the knowledge sources available, I can offer some relevant context here. The CORS configuration for your GoodData organization β€” including which headers are allowed β€” is managed on the GoodData side via your organization's CORS settings. The documentation confirms that allowed origins (and by extension, CORS behavior) are configured through: β€’ The Developer Settings UI on your GoodData instance (
https://analytics.golmn.com/settings
) β€’ Or via the
/api/v1/entities/admin/organizations/<organization_id>
API endpoint [Enable CORS for an Organization] However, the knowledge sources do not explicitly document the ability to configure
Access-Control-Allow-Headers
(i.e., which request headers are permitted in preflight responses) through the GoodData CORS settings UI or API. The documented CORS configuration covers allowed origins, not allowed headers. What This Means Practically The specific issue β€” that Datadog/RUM trace headers (
traceparent
,
x-datadog-trace-id
, etc.) are not included in the
Access-Control-Allow-Headers
response from
/features
β€” is likely something that would need to be addressed by GoodData's engineering team, as it relates to how the GoodData backend responds to preflight requests on that endpoint, not something you can configure yourself through the standard CORS settings. Recommended Action I'd strongly recommend raising this directly with GoodData Support, providing: β€’ The specific OPTIONS request/response headers from your PROD browser capture β€’ The list of headers being blocked (
traceparent
,
tracestate
,
x-datadog-*
) β€’ The endpoint affected (
/features
) This is beyond what can be self-configured based on the available documentation, and GoodData Support would be best positioned to confirm whether this is a backend change they need to make. πŸ‘ 0 button πŸ‘Ž 0 button ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups. πŸ‘€ Our Support team monitors all messages and will jump in if needed.
Based on the knowledge sources available, I can offer some relevant context here. The CORS configuration for GoodData Cloud is managed on the GoodData side β€” specifically, the
Access-Control-Allow-Headers
and
Access-Control-Allow-Origin
settings for your GoodData instance (e.g.,
analytics.golmn.com
) are controlled through GoodData's organization settings, not your own server. [Enable CORS for an Organization] However, the knowledge sources do not contain enough information to confirm whether GoodData's
/features
endpoint's
Access-Control-Allow-Headers
can be customized to include third-party headers like Datadog's
traceparent
,
x-datadog-trace-id
, etc. This is a very specific server-side CORS header configuration on GoodData's infrastructure, and that level of detail is not covered in the available documentation. What You Should Do This is a case where you need to contact GoodData Support directly and ask: > "Can the
Access-Control-Allow-Headers
for the
/features
endpoint on our GoodData instance be extended to include Datadog RUM/trace headers (
traceparent
,
tracestate
,
x-datadog-origin
,
x-datadog-parent-id
,
x-datadog-sampling-priority
,
x-datadog-trace-id
)?"
In the meantime, a practical workaround on your side would be to configure your Datadog RUM SDK to exclude the GoodData
/features
endpoint
from being instrumented with trace headers, which would prevent the preflight from including those headers in the first place. This is a Datadog-side configuration and would not require any changes from GoodData. πŸ‘ 0 button πŸ‘Ž 0 button ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups. πŸ‘€ Our Support team monitors all messages and will jump in if needed.
My answer remains consistent with what I provided previously: the knowledge sources do not document the ability to configure
Access-Control-Allow-Headers
(i.e., which request headers are permitted in preflight responses) through GoodData's self-service CORS settings. What GoodData's CORS configuration UI and API allow you to manage is allowed origins β€” not the set of headers the server will accept in preflight requests. [Enable CORS for an Organization] The
/features
endpoint's
Access-Control-Allow-Headers
response is controlled by GoodData's backend infrastructure, not by anything you can configure yourself through the Developer Settings. Your Options 1. Contact GoodData Support β€” Ask them to extend the
Access-Control-Allow-Headers
on the
/features
endpoint to include the Datadog/RUM trace headers. This is the most direct fix but requires action on GoodData's side. 2. Workaround on your side β€” Configure your Datadog RUM SDK to exclude the GoodData
/features
endpoint from distributed tracing instrumentation. This would prevent Datadog from injecting
traceparent
,
x-datadog-*
, etc. headers on that request, so the preflight would no longer include them and would pass. This is entirely within your control and does not require any changes from GoodData. Option 2 is likely the faster path to resolution while you engage GoodData Support on the longer-term fix. πŸ‘ 0 button πŸ‘Ž 0 button ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups. πŸ‘€ Our Support team monitors all messages and will jump in if needed.
My answer is the same as I've provided in the previous turns of this conversation, as the knowledge sources haven't changed. To summarize clearly: β€’ GoodData's self-service CORS configuration (via Developer Settings or the API) only controls allowed origins, not
Access-Control-Allow-Headers
. [Enable CORS for an Organization] β€’ The
Access-Control-Allow-Headers
returned by the
/features
endpoint is controlled by GoodData's backend infrastructure, and the knowledge sources do not document any way for you to modify this yourself. So yes β€” extending
Access-Control-Allow-Headers
on the
/features
endpoint to include the Datadog/RUM trace headers would be a change that needs to be made on GoodData's side, not yours. Practical Next Steps 1. Contact GoodData Support with your browser capture showing the OPTIONS request/response, and request that the Datadog trace headers (
traceparent
,
tracestate
,
x-datadog-origin
,
x-datadog-parent-id
,
x-datadog-sampling-priority
,
x-datadog-trace-id
) be added to the
Access-Control-Allow-Headers
for the
/features
endpoint. 2. In the meantime, configure your Datadog RUM SDK to exclude the GoodData
/features
endpoint from distributed tracing instrumentation β€” this is entirely within your control and would stop the problematic headers from being sent on that preflight, resolving the issue immediately without waiting on GoodData. πŸ‘ 0 button πŸ‘Ž 0 button ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups. πŸ‘€ Our Support team monitors all messages and will jump in if needed.
m
Hi Nick, I just wanted to let you know that we received your support ticket and we will follow-up with you directly there once we have reviewed this behaviour and have some clarification for you
gratitude thank you 1