Connecting concorbit to Autotask

Connecting Autotask gives concorbit read access to your PSA so engineers can work in concorbit while your existing Autotask records stay in lock-step. The connection is two-directional but defaults to pull-only. Every push surface is an opt-in toggle (see Sync toggles).

This page covers the one-off Connect step and the dispatch loop concorbit uses once connected. For the new-tenant checklist version of this, see the Setup checklist.

Prerequisites

You need an Autotask API user (not a regular Autotask user) with:

• A Security Level that can read Tickets, Contacts, Companies, Resources, Contracts, ContractServices, TimeEntries, and Projects.

• An API Tracking Identifier generated under Admin, Integrations, API Tracking Identifiers in Autotask. Give it a recognisable name like "concorbit" and save the integration code.

• A unique email and a saved Secret. The Secret is shown once at creation, save it before leaving the page.

Your Autotask administrator sets these up. concorbit never asks for a regular user's password.

Connect step

1. Open Settings, Autotask in concorbit (or /settings/autotask).

2. Paste the API user's email, the Secret, and the integration code into the Connect form.

3. Click Connect. concorbit:

   • looks up your Autotask zone (about one second),

   • validates the credentials with a small test query,

   • saves the encrypted credentials and marks the connection healthy.

4. If anything fails, the form tells you which step failed. The most common causes are a typo in the email or a wrong integration code.

A successful connect always starts the connection in mirror mode (the default lifecycle state). See Wind-down for what the other modes mean.

What gets pulled on first connection

Connecting alone does not back-fill data. The first sync is operator-triggered: once Connect has succeeded, configure your Mapping Master entries, then trigger the initial sync from the Autotask settings page. The orchestrator runs in dependency order, picklists first, then resources, companies, contacts, contracts, tickets, and the long-tail enrichments (notes, attachments, time entries, projects).

You can run a targeted sync for a single company at any time via the Sync from AT button on a company page, or via the CLI:

php artisan autotask:sync --entity=all --company=<concorbit-company-uuid>

The dispatch loop

Once steady-state, scheduled syncs run on a cadence:

• Tickets (with ticket notes pulled inline per ticket) every 5 minutes.

• Time entries every 15 minutes.

• Projects + tasks every 30 minutes.

• Contract blocks and attachments hourly (attachments also pull inline on every ticket pull, so a fresh ticket looks complete from its first sync).

• Invoices daily at 02:30, contract charges daily at 02:45 (invoices first so the mirror evidence they update already exists).

• Everything else — companies, contacts, contracts, resources, picklists, UDF definitions, services, contract lines, contract service units, adjustments, quotes — daily at 03:30 as part of the full re-sync. Contacts and contracts have no dedicated faster cadence; they only refresh on this daily sweep.

Each call goes through concorbit's per-provider rate limiter (5 requests per second for Autotask, well clear of the published ceiling). When Autotask returns 429 the API job releases itself with the Retry-After value and the call count increments on the platform API health page.

Disconnecting

Click Disconnect on Settings, Autotask. The connection is deactivated immediately. No more sync activity fires, your existing concorbit data stays, and on disconnect the connection mode resets to mirror so a later reconnect starts cleanly. If you intend to permanently retire the integration, prefer the wind-down ladder instead, which protects against double-billing and stranded revert windows.

Troubleshooting

See Sync toggles next for what flips push behaviour on and off.