Redo log can only tolerate storage error for 6s

[fubinzh]

What did you do?

1. Create CDC changefeed with redo log
2. Run workload
3. Inject network partition to redo log storage for 10min

What did you expect to see?

CDC should be able to tolerate 10min for redo log storage access error

What did you see instead?

CDC currently only tolerates storage error for 6s:

Versions of the cluster

bash-5.1# /cdc version
Release Version: v6.5.0-11-g3808d9732-dirty
Git Commit Hash: 3808d97
Git Branch: release-6.5
UTC Build Time: 2023-01-05 07:30:18
Go Version: go version go1.19.1 linux/amd64
Failpoint Build: false

[fubinzh]

/severity Moderate

[jebter]

/severity moderate