In the Helm Chart values.yaml we have several numebric values such as fullReconciliationIntervalMs or operationTimeoutMs. In some cases, when set to too big numbers, they seem to get misformatter as floats (e.g. 1800000 -> 1.8e+06). This breaks the Cluster Operator deployment as Java is not able to decode this:

Exception in thread "main" io.strimzi.operator.common.InvalidConfigurationException: Failed to parse. Value 1.8e+06 is not valid

This was originally raised on Slack in https://cloud-native.slack.com/archives/CMH3Q3SNP/p1754299547519069

While the reconciliation interval of 1800000 ms is not typicaly and not really recommended, we should try to fix it so that the number has always good format. Having it as a String already in values.yaml seems to help. But there seem to be also other options -> e.g. make sure it is interpreted/converted to int64 in the template (with something like value: {{ int64 .Values.fullReconciliationIntervalMs | quote }}) etc. Someone with better Helm knowledge should decide what the right format is the fix it accordingly.