Increased CortexIngesterReachingSeriesLimit critical alert threshold from 80% to 85%

[pracucci]

What this PR does:
The CortexIngesterReachingSeriesLimit alert currently has two threshold:

• warning alert: > 70%
• critical alert: > 80%

The warning alert is currently discussed in #362 while this PR is focusing on the critical alert.

What we've learned from production about ingesters reaching series limit:

• Series growth is typically slow and the alert fires way before we get even close to the limit (eg. in a Cortex cluster with 50 ingesters and limit of 2.5M / ingester, the ingesters in-memory series grow by 1% every 1.2M series)
• In case of emergency, it's very quick to increase the max series limit, since it's part of the runtime config (no ingesters restart is required)

Because of this, I would propose to fire the critical alert once we reach the 85% utilization (instead of the current 80% threshold). This will still leave us a 15% room before hitting the limit, which typically gives us enough time to address without any customer-facing disruption.

This change, together with the change we're doing in #362, should help to improve CortexIngesterReachingSeriesLimit alert to fire only when there's actually an action to take.

Which issue(s) this PR fixes:
N/A

Checklist

• CHANGELOG.md updated - the order of entries should be [CHANGE], [FEATURE], [ENHANCEMENT], [BUGFIX]