fix: incorrect uid in dashboard exists method

[weisdd]

In #1027, dashboard exists method was not updated to use correct uid. As a result, the operator will never stop reconciling dashboards with hardcoded uids creating a loop:

• the operator creates a dashboard and updates CR status with correct uid;
• since CR is updated, the operator receives an update via WATCH. That's where it was supposed to stop by checking that dashboard exists, hash is the same, and periodic resync timeout hasn't elapsed. Though, since uid is taken from CR instead of the dashboard, the operator would think the dashboard does not yet exist in grafana, so it reuploads it and updates lastResync in CR status;
• CR status update triggers another reconciliation, and so on.

Fixes: #1045