Hedera State Validator stability issues

State validation has been very unstable recently. Lots of failures, most are because of infrastructure or other issues, but they result in Rootly alerts anyway and must be fixed. The goal is to catch all real problem and filter out all false positives.

A few examples and comments:

* Request to https://status.hedera.com/api/v2/summary.json failed. Max attempts (5) exceeded

There should be more retries / longer waits between retries. Also, need to check with RE why github runners may have connection issues.

* Errors during initial setup for Previewnet Node N. Please check the job for details

The failing script is `.github/scripts/hedera-network-status.sh Previewnet "false"`, so it may be related to the previous one.

* Entity ID counter mismatch, e.g. EVM hook storage counts or token rel counts. There is a ticket about it filed some time ago, but the problem has become more visible since then: https://github.com/hiero-ledger/hiero-consensus-node/issues/24291

* Diff test failed for node N (round XYZ)

This seems to be a problem with downloading block streams, but I may be wrong. A workaround has been applied for this issue, but the underlying problem should still be identified and resolved.

* Migration test failed for node N (round XYZ)

Java version mismatch?

```
Error: LinkageError occurred while loading main class com.hedera.statevalidation.StateOperatorCommand
java.lang.UnsupportedClassVersionError: com/hedera/statevalidation/StateOperatorCommand has been compiled by a more recent version of the Java Runtime (class file version 69.0), this version of the Java Runtime only recognizes class file versions up to 65.0
```