Now we start a background job that waits forever for the NodeMetrics model to be synced (using exponential backoff, maxing out at checking once per minute). If the model never becomes available, that job will keep checking for the Model once per minute, and will log an error once every 5 minutes. If the model ever becomes available, however, it will start up the NodeMetrics publishing subsystem at that point and the background task stops.
New approach instead of https://github.com/ceramicnetwork/js-ceramic/pull/3265.
Now we start a background job that waits forever for the NodeMetrics model to be synced (using exponential backoff, maxing out at checking once per minute). If the model never becomes available, that job will keep checking for the Model once per minute, and will log an error once every 5 minutes. If the model ever becomes available, however, it will start up the NodeMetrics publishing subsystem at that point and the background task stops.