Queue storage error

[jcw-]

Describe the bug
When doing a /load query after the refreshKey expiry has been reached (e.g. 5 minutes), the following error is reported in the logs:

{"level":30,"time":1582999195218,"pid":40625,"hostname":"Jcs-MBP-2","req":{"id":1,"method":"GET","url":"/cubejs-api/v1/load?query=%7B%22measures%22:%20[%22Test.amount%22],%22timeDimensions%22:%20[%7B%22dimension%22:%20%22Test.createdAt%22,%22granularity%22:%20%22day%22,%22dateRange%22:%20%22Last%20year%22%7D],%22dimensions%22:%20[],%22filters%22:%20[]%7D","remoteAddress":"::1","remotePort":58172},"timings":{"Load Request SQL":590,"Query Execution":175},"res":{"statusCode":200},"responseTime":778,"msg":"request completed","v":1}
{"message":"Queue storage error","queryKey":"b5f2ddc43b3c499c764272a3ea59b2b8","error":"Error: Can't update b5f2ddc43b3c499c764272a3ea59b2b8 with {\"startQueryTime\":1582999195274}\n    at RedisQueueDriverConnection.optimisticQueryUpdate (/Users/jcw/workspace/cube.js/packages/cubejs-query-orchestrator/orchestrator/RedisQueueDriver.js:145:11)\n    at process._tickCallback (internal/process/next_tick.js:68:7)","queuePrefix":"SQL_QUERY_STANDALONE_default"}

This is a query hitting an external rollup defined like this:

  preAggregations: {
    categoryAndDate: {
      type: `rollup`,
      measureReferences: [Test.count, Test.amount],
      dimensionReferences: [bar, foo, Accounts.name],
      timeDimensionReference: createdAt,
      granularity: `day`,
      partitionGranularity: `month`,
      refreshKey: {
        every: `1 day`,
        incremental: true,
        updateWindow: `12 month`
      },
      external: true
    }
  }

To Reproduce
Steps to reproduce the behavior:

1. Define a rollup
2. Query the rollup via /load
3. Not sure if you see it first time, but if you wait for 5 minutes the next /load will trigger the error in the logs.

Expected behavior
No errors

Version:
master branch (0.17.1)

Additional context
The exact log format may differ from what is show above.