COS / Alerts MongodbReplicationLag and MongodbReplicationHeadroom with errors

## Steps to reproduce
Follow COS integration tutorial from Charmhub
Use channel 6/edge for Mongodb

## Expected behavior
Not to see alerts with errors

## Actual behavior
MongodbReplicationLag and MongodbReplicationHeadroom alerts with errors


## Versions
```bash
# (base) blue@blue:~$ juju status
Model Controller Cloud/Region Version SLA Timestamp
mongodb lxd localhost/localhost 3.5.4 unsupported 22:24:34Z

SAAS Status Store URL
alertmanager-karma-dashboard active lxd admin/cos.alertmanager-karma-dashboard
grafana-dashboards active lxd admin/cos.grafana-dashboards
loki-logging active lxd admin/cos.loki-logging
prometheus-receive-remote-write active lxd admin/cos.prometheus-receive-remote-write

App Version Status Scale Charm Channel Rev Exposed Message
grafana-agent active 3 grafana-agent latest/stable 223 no
mongodb active 3 mongodb 6/edge 206 no

Unit Workload Agent Machine Public address Ports Message
mongodb/0 active idle 0 10.58.159.207 27017/tcp
grafana-agent/1 active idle 10.58.159.207
mongodb/1 active idle 1 10.58.159.88 27017/tcp
grafana-agent/2 active idle 10.58.159.88
mongodb/2* active idle 2 10.58.159.249 27017/tcp Primary
grafana-agent/0* active idle 10.58.159.249

Machine State Address Inst id Base AZ Message
0 started 10.58.159.207 juju-0c41a2-0 ubuntu@22.04 Running
1 started 10.58.159.88 juju-0c41a2-1 ubuntu@22.04 Running
2 started 10.58.159.249 juju-0c41a2-2 ubuntu@22.04 Running
```

Reproduce the issue by following the COS integration tutorial from Charmhub with MongoDB on channel 6/edge and the three-unit setup shown in the report. Inspect the MongodbReplicationLag and MongodbReplicationHeadroom alert errors and compare them with the expected alert behavior. Done means the alerts no longer report errors.

Written by the indexing model from the issue text.