Zendesk

The zendesk connector reads Zendesk Support tickets and their audit history.

#What it pulls

Records. One ticket record per ticket updated in the window: ID, created and updated times, type, subject, description, priority, status, requester, submitter and assignee IDs, organization, group and brand IDs, channel, tags, satisfaction score and due date.

Episodes. For each of those tickets, its audits: the values set when the ticket was created, every field change with its old and new value, public comments and internal notes, each with its time and author. Requester, submitter and assignee changes are recorded by the person's name, which the scrubber turns into a token.

Subjects, descriptions and comment bodies are free text, which is redacted. Support tickets often name customers in the body, so read the report's free-text notes, and consider the optional entity detector. See How scrubbing works.

#Credentials

1. In Zendesk, go to Admin Center → Apps and integrations → Zendesk API and turn on token access.
2. Add an API token.
3. Give the SDK your subdomain (acme for acme.zendesk.com), the email of an admin agent and the apiToken.

You can pass an OAuth accessToken with the read scope instead of the email and token. datayield connect zendesk fails if the credentials are not recognized, or if the user is not an admin.

#Connect

datayield connect zendesk -o subdomain=acme -o email=admin@acme.example -o apiToken=env:ZENDESK_TOKEN

Secret options passed directly are saved to ~/.datayield/credentials. Values written as env:NAME are read from the environment at run time, which suits CI. See CLI reference.

#Options

#New records on each run

The connector uses Zendesk's cursor-based incremental ticket export, starting from the last successful run, and keeps tickets whose update time falls in the window. A ticket updated again later is sent again with its new state, and its episode is rebuilt from the full audit history.