Amazon Lookout for Equipment is no longer open to new customers.
Existing customers can continue to use the service as normal.
For capabilities similar to Amazon Lookout for Equipment see our
blog post
Understanding the inference process
When you're planning your use of Lookout for Equipment, it may be useful to understand exactly what happens at each step of the inference process.
Understanding inference scheduling windows
When you schedule inference, you may set your data upload frequency time to any of the following values, in minutes: 5, 10, 15, 30, 60.
Lookout for Equipment then calculates the base number of segments per hour by dividing 60 by the length of your segments.
You may also set an offset window in increments of minutes, from 0 to 60.
At the beginning of each segment, Lookout for Equipment waits for the offset window to close before running inference.
At the top of the hour, the process begins again.
| Inference interval | Inferences per hour | First inference after 09:00 (with no offset) | First inference after 09:00 (with a 5-minute offset) |
|---|---|---|---|
|
5 |
12 |
09:05 |
09:10 |
|
10 |
6 |
09:10 |
09:15 |
|
15 |
4 |
09:15 |
09:20 |
|
30 |
2 |
09:30 |
09:35 |
|
60 |
1 |
10:00 |
10:05 |
The inference process
-
Lookout for Equipment looks for the component name (which can be the name of an asset or a sensor, depending on how your data was ingested).
-
Once the component name is found in the file name, Lookout for Equipment looks at the time stamp in the CSV file name.
-
The timestamp in the file name must be within the range of time that your scheduler is running. For example, if the scheduler is running every 5 minutes, then at 9:05, Lookout for Equipment will look for any files that have a timestamp from 9:00 to 9:05. Any files with timestamps outside this range will be ignored for the inference run.
-
Lookout for Equipment automatically ingests the files with the right component name, and within the right time range.
-
Lookout for Equipment opens the CSV file and runs inference on any rows in the CSV file with timestamps that fit within the scheduler window. For example, if the scheduler is running every 5 minutes, and the current time is 9:05, then Lookout for Equipment will grab any files with the timestamp in the file name from 9:00 to 9:05, and will then run inference on any rows in the CSV with timestamps between 9:00 to 9:05.
-
The inference results are placed into your designated output bucket in a JSON file.
-
The steps above are repeated in perpetuity until the scheduler is turned off.
