Many DBSS Usability Connection Availability false positive alarms at the same time (4381090)

The DBSS Usability Connection Availability alarm indicates that the SQL Server agent was unable to establish a JDBC connection to the monitored instance. The alarm message typically includes:

• TCP/IP connection failed or Timeout occurred: 60 seconds.
• Profile details such as host, instance, port, and authentication type.

These alarms are triggered en masse when numerous agents cannot validate database usability within the configured timeout.

Common causes include:

1. Network Connectivity Issues

   • Temporary latency or packet loss between the FglAM and the SQL Server hosts.
   • Firewall or IDS/IPS inspection delaying TCP handshakes.

2. FglAM Resource Contention

   • Thread pool saturation or JVM pauses causing delayed agent tasks.
   • High concurrency from multiple agents on the same FglAM.

3. SQL Server Unavailability

   • Service stopped or restarted during the collection window.
   • Instance not accepting TCP/IP connections.

1). Validate SQL Server Availability

• Confirm the SQL service was running during the alarm window.
• Check SQL error logs for downtime or connectivity errors.

2). Check Network Path

• Test connectivity from the FglAM host to the SQL Server (ping, telnet to port 1433).
• Review firewall rules and network monitoring for latency spikes.

   

3). Review FglAM Health

• Check FglAM logs for thread pool delays or JVM GC pauses.
• Ensure FglAM has sufficient resources (CPU, memory).

4). Preventive Measures

• Stagger agent schedules to reduce concurrency by adding this JVM parameter to the baseline.jvmargs.config file and restarting the FglAM:
  vmparameter.0 = "-Dagent.collector.schedule.load.max.delay.millis=300000";

This spreads agent activation over 5 minutes instead of the default 2 minutes, reducing collection overlap.

• Increase socketTimeout if network latency is high.
• Consider deploying additional FglAMs closer to DB servers. 

Note

• These alarms are not false positives if the agent genuinely failed to connect.
• If connectivity recovered automatically, the issue was likely transient (network or FglAM load).