Reduce recoverer workload

[ferglor]

https://smartcontract-it.atlassian.net/browse/AUTO-10074

• Bump to a version of automation with reduced workload/cadence for recoverer
• Reduce cached logs in core

Performance comparison

CL node stats

Enabled

Disabled

Results

When enabled, CPU usage peaks at over 2.5 and is quite erratic, mem usage approaching 2.5GB When disabled, CPU peaks at just over 1.0 with occasional spikes, memory spike to 2GB but mostly stays at 1.25GB

DB node stats

Enabled

Disabled

Results

When enabled, CPU peaks at around 1.0, with early RAM spikes to 1 GB, but doesn't go much higher than 1GB overall When disabled, CPU peaks just over 0.3, with no RAM spikes, but overall higher usage of 1.25

Missed logs

Enabled

Disabled

Results

No stats are reported when the recoverer is disabled