Listed below are issues and solutions you may encounter in ArcGIS Analytics for IoT when ingesting event data. If you encounter an issue not addressed below, work with your administrator to contact Esri Support.
For issues and solutions related to accessing the Analytics for IoT application, signing in, or creating and starting items, see Troubleshooting for general issues and solutions.
Data validation errors
If using the Website (Poll) or RSS feed or data source, ensure the website response or file being polled does not exceed 8 MB.
For other feed or data source types, the time it takes to sample data may exceed the timeout for sampling. This may be due to the data being too large or there being not enough samples received during the sampling period. Either try resampling or use the workflow to manually enter samples.
It is important to ensure that the schema defined on the feed or data source matches the data coming in and the data is externally accessible.
- Did the samples retrieved on the feed or data source configuration fully encompass all possible attributes?
- Did you provide samples that do not actually match the incoming data?
- Did the format of the data change since the feed or source was configured?
- Is the geometry configured correctly?
- Is the external system or resource still accessible over the internet with the defined connection properties?
Ensure Analytics for IoT can still connect to the external system or resource (broker, website, cloud IoT resource, or other). Try editing the feed or data source and resampling data as the configuration process checks for connectivity. Creating a new feed or source may be faster than troubleshooting a pre-existing one. Note, if the schema changes this may affect any analytic outputs that use the feed or data source and they may also need to be recreated.
If the problem is with a feed that polls periodically for data, also consider the frequency of the polling schedule. Polling occurs on regular clock intervals as opposed to the start time of the feed. For example, if a feed is set to poll for data every 5 minutes, and the feed is started at 8:03am, the first request for data will occur at 8:05am, the next at 8:10am, and so on. For data that changes infrequently, a recommended best practice is to test the feed with a polling interval of every 10 or 30 seconds to ensure data is successfully retrieved, and then update the feed with the appropriate polling interval.
Are any field types incorrectly defined? For example:
- Defining a field as numeric type when in fact string character values are also present in the data associated with that field will result in no data being ingested.
- Setting a Float64 field as a Float32 field would alter data values by dropping precision.
For date fields created from strings, ensure the data formatting string is accurate for all incoming date values.