optimize(op-node): continue optimizing sequencer step schedule

[bnoieh]

Description

continue optimizing sequencer step schedule

Rationale

there're still some corner cases, in which sequencer block produce may has issue:

1. if whole l1 network went down for more than maxSequencerDrift time, sequencer was unable to auto-recover after l1 network recovered
2. when high l1 network latency last for a long time(> maxSequencerDrift), sequencer will produce block but difficult to produce full block
3. when sequencer step is hot(e.g. catching up), admin_stopSequencer operation may failed due to context timeout

this PR will fix above cases.

Example

n/a

Changes

1. call sequencerStep() immediately rather than reset sequencerTimer
2. handle l1HeadSig before call sequencerStep() to avoid error caused by cannot find next l1 block
3. add accEmptyBlocks in sequencer.go to avoid never produce full block in corner case
4. add more steps to high priority schedule