PMS Sync Auto-Retry After Error

Available from: v0.1.281

Overview

When a Property Management System (PMS) connection sync fails, the platform now automatically schedules a single retry attempt 30 minutes after the failure is detected. This reduces manual intervention for transient errors — such as brief network outages or temporary third-party API unavailability — without risking runaway retry loops.

How It Works

The workflow is event-driven and operates as follows:

1. Trigger

The workflow listens for the pms/connection.sync.failed event. This event fires when a pmsConnections record transitions its status field to error.

2. Wait Period

Upon receiving the event, the workflow pauses for 30 minutes before taking any action. This delay is intentional — it provides a window for transient issues (e.g. third-party downtime, network blips) to resolve themselves.

3. Single Retry Attempt

After the wait, the workflow re-queues one sync attempt for the affected PMS connection. The attempt is logged to pmsSyncLogs.

4. Outcome Handling

If the retry succeeds: The pmsConnections status is updated accordingly and normal sync cadence resumes.
If the retry also fails: The auto-retry workflow takes no further action. The existing pms-sync-failure-notify workflow handles escalation (notifications, alerting, etc.).

Affected Entities

Entity	Role
`pmsConnections`	Source of the `status: error` transition that triggers the workflow
`pmsSyncLogs`	Records the retry sync attempt and its outcome

Retry Policy

This workflow enforces a maximum of one automated retry per failure event. This is by design:

Prevents retry storms against third-party PMS APIs.
Keeps the retry window predictable (30-minute delay, single attempt).
Defers to human-reviewed escalation for persistent failures.

If more than one retry is needed, it must be initiated manually or through the escalation workflow.

Escalation

Persistent sync failures — where both the original sync and the auto-retry fail — are handled by the pms-sync-failure-notify workflow. This pre-existing workflow is responsible for sending notifications and alerts to the relevant parties. The auto-retry workflow does not duplicate or replace this behaviour.

Summary Flow

pms/connection.sync.failed
        │
        ▼
  [Wait 30 minutes]
        │
        ▼
  [Queue retry sync]
        │
     ┌──┴──┐
  Success  Failure
     │        │
  Resume   pms-sync-failure-notify
  normal    handles escalation
  sync