forcedotcom / scmt-server Goto Github PK

Truncate the field
or make the domains field longer?

Ensure that the web browser's XSS filter is enabled, by setting the X-XSS-Protection HTTP response header to '1'.

Other information

The X-XSS-Protection HTTP response header allows the web server to enable or disable the web browser's XSS protection mechanism. The following values would attempt to enable it:

X-XSS-Protection: 1; mode=block

X-XSS-Protection: 1; report=http://www.example.com/xss

The following values would disable it:

X-XSS-Protection: 0

The X-XSS-Protection HTTP response header is currently supported on Internet Explorer, Chrome and Safari (WebKit).

Note that this alert is only raised if the response body could potentially contain an XSS payload (with a text-based content type, with a non-zero length).

Sending an email to an alias so that we can monitor who is using the tool/package

If I select migrate data range for accounts, I don't get a different search url. Here's an example:

https://zzz-minjunyang.desk.com/api/v2/companies/search?q=(created_at:[1%20TO%201498525952])&per_page=500&page=1&sort_field=created_at&sort_direction=asc

The minimum created at date here is 1, but it should be the date I select in the UI.

Some cases have missing interactions, please see an example here: http://screencast.com/t/PKjSkg3M

Specify which objects you would like to migrate and then migrate them in succession.

Big blocker here is how to handle failures of a parent object.

Currently all environmental variables are required when deploying from the Github project. That's fine for most of them but PACKAGE_NAMESPACE can be an empty string and should not be required.

Investigation is needed to answer some questions like is the variable going to be set if not required and value is an empty string?

Domains field doesn't correspond to default Account field (Currently it's mapped to website, but Desk allows multiple domains). We can potentially put all domains as comma separated in a custom field.

Deleted cases are migrated as resolved if they are not permanently deleted in Desk.com, the tool should filter out the deleted cases when doing the migration.

Cases were migrated as "Web" for origin and "Phone" for channel for QNA cases.

Quoted email content is not migrated. Need to add HtmlBody and Headers for email messages.

Update Readme with License information

Dedup process (i.e using external ID, email, etc.) is a nice-to-have.
-Currently, there is no de-duplication process. Accounts with the same name or contacts with the same email address will be created more than once.

Workaround the API rate limit

(so if there's multiple migrations running and they all hiccup from the same perf issue they don't all slam us at the same time?)

Title and background are not migrated for contacts

"ClosedDate" should be the actual last updated date from Desk, currently it's the migration date. (Need to do more research on SF side, might not be changeable)

Only one email address by default (this might be best managed with field mappings and custom fields).

If Desk returns an email in the format "[email protected]" [email protected], then it will fail when translated to Salesforce.

Scenario: I'd like to migrate my email, notification and Auto-Acknowledgment themes from Desk

Desk allows duplicate picklist values for custom fields. But Salesforce doesn't. We should dedup the values before creating the custom fields in Salesforce.

The legacy code is completely untested, in order to build confidence in the existing code base and to make sure functionality stays the same we need to create test cases.

The content of app.json is not valid; please see app.json Schema for more information.

• name: is too long (maximum is 30 characters)
• description: is too long (maximum is 140 characters)

We need to migrate inlined images and rehost them in SFDC.

1. Scan article body for desk hosted images
2. Upload these images to Salesforce, retrieve new url
3. Replace the links in the article body
4. Upload article

Remove whitespace from consumer key/token and access key/token during step 2

Right now, running an interaction migration creates multiple bulk import jobs, so the records total and records failed count is not there for phase 2 release, will work on this as a phase 3 item.

Customer asked if there was a way to see the progress on the server - maybe it can be exposed somehow? Might clash with ToS of the papertrail heroku add on we're currently using.

CSat data is not migrated.

UI issue in calendar picker (using chrome latest version): http://screencast.com/t/uZUc4bfxf1b

Timestamps are not migrated correctly for replies and notes for some cases.

Many of our customers have Salesforce implemented already and have existing contact/account records that match up with Desk records.

1. Determine how the tool currently handles duplicates
2. Determine what strategies are available to mitigate duplicate creation

We're missing two environmental variables in app.json:

SALESFORCE_TRACE_MESSAGES = 1
JAVA_ENV = development

Generate a CSV file of all the error logs on each object with detailed error message

@minjunyang88 commented on Mon Mar 06 2017

Ensure that the migration user has R/W permission for the “unassigned” queue.

@tstachl commented on Thu May 18 2017

@minjunyang88 does this not happen during the "Group Member" migration?

@minjunyang88 commented on Thu May 18 2017

No. I had to manually assign myself to the "unassigned" queue.

Desk articles can have attachments. We'll likely need to add a file field to SFDC to contain these because it doesn't exist by default.

In order for the jobs to work with an unmanaged package we load the namespace for the managed package from an environmental variable during the slug compilation. The app.json file needs to have the environmental variable specified for this process to work.

Currently we can only migrate full data sets or partial data sets. This makes it difficult to deal with failed uploads. We need a way to identify and retry these.

Both Desk KB and SFDC Knowledge allow translated articles. Currently we only migrateEenglish language articles as specified by the accept language header.

1. Retrieve Desk enabled languages. We already have a method for this (In Desk Service, getDeskSiteLanguagesMap())
2. Retrieve SFDC Knowledge enabled languages.
   -https://developer.salesforce.com/docs/atlas.en-us.api_meta.meta/api_meta/meta_knowledgesettings.htm
3. Map them and determine which languages should be retrieved from Desk. Not all the languages will be able to map over because SFDC supports fewer languages.
4. Upload articles

Header would look like X-SC-MIGRATION=[site_id]

When user clicks the button to migrate attachments, there is no feedback/popup like other steps. We should look into add the the pop up so that user doesn't click twice, which will actually migrate duplicate attachments.

If a batch is greater than 10M characters, the batch fails and the whole migration stops. Also, the job is left running and never actually closes.

We should rescue this error for at least interactions and make the batch size incrementally smaller.

need to check sfdc limit

In order to use the unmanaged code from the forcedotcom/scmt-app we need to allow for not having a namespace.

Remove company filter option from UI since we don't allow that search logic in Desk for company object.

Assigned agents that are deleted shouldn't be migrated.

Don't migrate deleted users, cases with deleted users would be unassigned on the case level after migration.