Connect Autotask and choose what syncs

Autotask is one of Current's three first-class PSAs (alongside ConnectWise and HaloPSA): a durable, two-way sync that keeps Autotask the source of truth for time and billing while Current owns scheduling and the views your team lives in.

• ＋An Autotask API user (not a regular login) with access to Projects, Tasks, Tickets, Resources, and Time Entries.
• ＋Tenant Admin role in Current — integrations are an admin-only action.
• ＋Optional but recommended: sandbox credentials for a dry run.

1. 1
   Open Settings → Integrations → Autotask

   Only Tenant Admins see this page.
2. 2
   Enter the API user's credentials

   Username, secret, and tracking identifier. Current discovers your Autotask zone automatically from the username — you never enter a zone URL.
3. 3
   Run the test connection

   Current validates the credentials against Autotask before storing them. Secrets are encrypted at rest and never sent to the browser again.
4. 4
   Set the poll interval

   Default is every 5 minutes. You can lengthen it per workspace if your Autotask API usage is near its threshold. Each workspace syncs in its own worker with its own time budget, so another workspace's heavy sync never slows yours down.
5. 5
   Start the initial pull

   Companies, contacts, projects, phases, tasks, tickets, resources, and time entries come across. Large books take a few hours; the sync resumes automatically if interrupted.

The field-ownership map prevents the two systems from fighting. Current never overwrites an Autotask-owned field on push, and Autotask pulls never clobber a Current-owned field.

• ＋Polling plus an optional webhook backstop — changes land within one poll cycle.
• ＋A durable job queue: every sync is a job that retries with backoff (up to 5 attempts) and dead-letters if it keeps failing, so nothing is silently lost.
• ＋Idempotent writes: running the same job twice never creates duplicates.
• ＋Outbound writes are confirmed by read-back — after pushing a time entry, Current reads it back from Autotask to confirm the approval status.

The Per-entity table shows one row per record type with a colored status. Synced (green) means it's fully caught up. Catching up (blue) means the sync is mid-walk and making progress under the sync watchdog's coverage — it synced what it could inside this cycle's time and request budget and continues on the next one, which is completely normal for a large book. That same in-progress row turns amber when the watchdog flags it: no forward progress for about 2 hours, or falling behind its expected refresh window (as soon as an hour for core project data; longer for slower-moving record types). A few low-stakes record types (like holidays and time off) sit outside watchdog coverage, so an in-progress cycle there shows plain amber Partial rather than blue. Amber means it's worth a look, not that anything is broken. Red is a genuine failure. The note under a status always spells out exactly what happened.

At the top of the page, an all-systems summary chip gives you the one-glance answer. It reads All systems green when every record type is healthy (and tells you how many are still catching up, if any); amber when something needs attention (a warning-level alert, a record type behind, or a partial the watchdog has flagged); red when there's a real issue (a failed job, a lost connection, or a sync that's stalled); and a grey Checking… while it's still loading the facts. It never shows green until it's sure — if anything is still unknown, it stays grey rather than claim everything's fine.