More monitoring SLIs tuning

Following up on !4730 (merged):

• Update the thanos-store SLI to monitor all unary GPRC requests, because why shouldn't we and it looks stable
• Drop the Grafana SLIs apdex score to 0.92 while we work on improving the Thanos datasource query latency. Unfortunately grafana_http_request_duration_seconds_bucket doesn't go above 5s so we can't raise satisfiedThreshold any higher and have to lower the apdex score instead.