Improvement feedback new APP SaaS Upgrade Assistant

[MarcioKaue]

We approved the APP for our Managed to SaaS migration case and below are the improvement points we identified (out of 15k configurations, an error occurred in 5k):

object ID = builtinmetric.metadata

Many errors related to Cloud or Extensions built-in metrics, wouldn't it make sense to skip this type of configuration?

--------------------------------------------------------------------------------------------------------------

objects ID = builtinspan-attribute and builtinspan-event-attribute

message example:

Schema "builtin:span-attribute" is deprecated - please update your configurations: this setting was replaced by 'builtin:attribute-allow-list' and 'builtin:attribute-masking

This seems like a problem to be fixed where the JSON provided in the API's GET is not accepted to make the POST because the object's id has been "updated"

-----------------------------------------------------------------------------------------------------------------------------

objects ID = builtinsynthetic.browser.assigned-applications, builtinsynthetic.browser.outage-handling, synthetic-monitor, builtinsynthetic.browser.name, builtinsynthetic.http.scheduling, builtinsynthetic.http.performance-thresholds, builtinsynthetic.http.cookies

message example:

Skipping deployment of toDeploy:builtin:synthetic.browser.assigned-applications:{ID}, as it depends on toDeploy:synthetic-monitor:SYNTHETIC_TEST-{ID} which was not deployed after root dependency configuration toDeploy:synthetic-location:SYNTHETIC_LOCATION-{ID} failed to deploy

We have environments with 200 to 900 synthetic monitors, it is completely unfeasible to update the ID property manually 1 by 1, this is a point of attention that needs to be evolved (applies to any similar dependency error by entity ID of the env origin).

--------------------------------------------------------------------------------------------------------------------------

object ID = calculated-metrics-service

message:

failed after 60 retries: (HTTP 400) (HTTP 400): {"error":{"code":400,"message":"Constraints violated.","constraintViolations":[{"path":"conditions","message":"At least one condition of the following types must be used: [SERVICE_DISPLAY_NAME, SERVICE_PUBLIC_DOMAIN_NAME, SERVICE_WEB_APPLICATION_ID, SERVICE_WEB_CONTEXT_ROOT, SERVICE_WEB_SERVER_NAME, SERVICE_WEB_SERVICE_NAME, SERVICE_WEB_SERVICE_NAMESPACE, REMOTE_SERVICE_NAME, REMOTE_ENDPOINT, AZURE_FUNCTIONS_SITE_NAME, AZURE_FUNCTIONS_FUNCTION_NAME, CTG_GATEWAY_URL, CTG_SERVER_NAME, ACTOR_SYSTEM, ESB_APPLICATION_NAME, SERVICE_TAG, SERVICE_TYPE, PROCESS_GROUP_TAG, PROCESS_GROUP_NAME]","parameterLocation":"PAYLOAD_BODY","location":null}]}}

I understand it as a product failure, since the Calculated Metric is normally created with the attribute condition as shown in the JSON:

attribute":string"REQUEST_TYPE"

And the Dynatrace API doesn't support it.

------------------------------------------------------------------------------------------------------------------

object ID = request-attributes

message: Deployment failed - Dynatrace API rejected HTTP request / JSON data: Failed to create DT object pix (HTTP 403)! Response was: {"error":{"code":403,"message":"Token is missing required scope. Use one of: CaptureRequestData (Capture request data)"}} (HTTP 403): {"error":{"code":403,"message":"Token is missing required scope. Use one of: CaptureRequestData (Capture request data)"}}

The documentation regarding Token creation needs to be updated, as it does not request this scope!

"If you decide to manually create a token it has to be scoped with:

Write configuration
Read configuration
Write settings
Read settings

Write SLO
Read SLO
Access problem and event feed, metrics, and topology
Create and read synthetic monitors, locations, and nodes"

[Radoslaw_Szulgo]

Thanks for feedback! We'll definitely try to improve on all you suggested in upcoming version release.

[erh_inetum]

Hi @Radoslaw_Szulgo ,

We are trying to upgrade the configuration with the tool.

We've uploaded the data sucessfully and the tool recognizes 108 config types and 1913 config objetcs. 

We click on Run and nothing happens.

On the logs we see this:

"Error: Deployment failed - check logs for details: failed to get independently sorted configs for environment \"XXXXXX\": There are 3 dependency cycles between the configurations.\nPlease check the following configuration's references and break the cycle:\n\"toDeploy:dashboard-share-settings

We have checked that nothing was migrated.

We know some dashboards have dependencies to an elements that don´t exist anymore. 

Could it be the reason if the fail? Is it normal that nothing be upgraded due to dependencies related to dashboard? How can we fix it?

Thanks in advance.

Regards,

Elena.

[Radoslaw_Szulgo]

Automatic resolving of cyclic dependencies are not yet supported. In this case you need to go to your Managed environment and see that dashboard share settings. Fix it. Then export the configuration and you can use the SaaS Upgrade Assistant. 

[erh_inetum]

Hi @Radoslaw_Szulgo ,

Another question: is it possible to exclude config types from the upgrade or it's not possible and all config types must be uploaded?

Thanks again,

Elena.

[Radoslaw_Szulgo]

Excluding is possible in version 1.0.1 This should be available to you today. 

[erh_inetum]

Hi @Radoslaw_Szulgo ,

Thanks a lot for your answers.

Onces we fixed the cyclic dependencies we got deploy the configuration on SaaS successfully.

I've created this product idea to the tools allow selecting only some config types. 

Thus, the tool could cover the two approaches: big bang and phase.

Thanks a lot.

Regards,

Elena.