MongoDB alarm Differences: Server State Change vs Availability (4380507)

This article explains the difference between the MongoDB Server State Change and MongoDB Availability alerts in Foglight. It also covers why only the state change alert fired when your MongoDB host was down for a few hours.

• MongoDB Server State Change: This alert triggers when a member of a MongoDB replica set changes its state, such as moving from PRIMARY to SECONDARY or to DOWN. It checks if the server's state is different from the last time it was checked.
• MongoDB Availability: This alert is designed to trigger if a monitored MongoDB server is unreachable two times in a row. It checks for network or process failures and expects the server to be unavailable for at least two collection cycles (about 10 minutes).
• If Foglight cannot reach the host at all (such as during a crash or network issue), the system may not trigger the availability alert because it loses contact and cannot confirm the status.

• You may only see the state change alert if Foglight was still able to contact the MongoDB instance and detect its state changed.
• The availability alert needs to see two failed collections in a row. If the host becomes "missing" or "stale" in Foglight, the alert may not trigger.
• To improve coverage, consider:
  • Tuning alert thresholds and logic.
  • Enabling or adjusting "object missing" alerts for when hosts disappear entirely.
  • Making sure monitoring agents handle unreachable hosts robustly.

Summary Table