During the Helm Hart upgrade from v0.9.4 to 0.9.5 the following steps need to executed additionally to the following guide:
The Kafka StatefulSet has now a startupProbe to work with long startup times. For this you need to change the following values in the values.yaml:
startupProbe: enabled: true failureThreshold: 600 periodSeconds: 10 timeoutSeconds: 10
Minor API changes to availability and performance, which will both behave similar to OEE now and return multiple values for longer time ranges. In Grafana panels using them must be change to use not the “last” calculation, but the “mean” calculation.
If you are using the modifyState message, ensure that your messages use the “timestamp_ms” & “timestamp_ms_end” notion, instead of the old “start_time_stamp” & “end_time_stamp” (https://github.com/united-manufacturing-hub/united-manufacturing-hub/pull/1137/files)
If you where using “deleteShiftByAssetIdAndBeginTimestamp” or “deleteShiftById”, please use “deleteshift” instead
If you where using “modifyProducedPieces”, please ensure that you now also send a timestamp_ms