Investigate QN Degradation Alerts from Checkly

## Context

We should look into those degradations. In general that's something I've seen since long ago, sometimes QN will take up to 5 seconds to respond to queries. Atlas fires a "burst" of queries when it loads - 7 at startup and then couple more later. If the QN is in this degraded state, all of the queries hang for a couple seconds and then usually resolve at the same time, suggesting that QN is hanging somehow. This could explain random long loads in Atlas

## Scope
- [ ] Identify the root cause of the frequent degradations (daily)
- [ ] Scope and raise an issue to fix the problem