This documentation contains a list of metrics and alerts that are available when monitoring PostgreSQL on Windows instances. Send us feedback if there are additional metrics or alerts you'd find valuable.

Redgate Monitor collects host machine data from Windows over WMI. You must use a Base Monitor installed on Windows to monitor PostgreSQL on a Windows host. In Redgate Monitor SaaS, you must use an agent installed on Windows. A Linux Base Monitor cannot collect data from Windows host machines.

Metrics

The following lists the metrics Redgate Monitor collects and exposes on the Analysis page:

Host Windows machine metrics

• Machine: processor time
• Avg. CPU queue length
• Machine: memory used
• Memory pages/sec
• Network utilization
• Disk used bytes
• Disk used %
• Disk free bytes
• Disk avg. read time
• Disk avg. write time
• Disk read bytes/sec
• Disk write bytes/sec
• Disk throughput bytes/sec
• Disk transfers/sec
• Avg. disk queue length

VMware metrics

Available when the Windows host machine is a VMware virtual machine:

• Processor time
• Guest CPU ready time
• Guest CPU swap wait
• CPU Co-Stop
• Memory swap in rate bytes/sec
• Memory swap out rate bytes/sec
• Memory usage
• Guest Memory latency
• Guest Memory ballooned
• Host Memory ballooned
• Disk latency
• Disk read bytes/sec
• Disk write bytes/sec
• Disk throughput bytes/sec
• Network usage bytes/sec
• Effective VM speed MHz

PostgreSQL on Windows metrics

Available at the instance level as an aggregate, and at the individual database level:

• Number of connections
• Transactions committed / rolled back
• Query overflow to temp data bytes sec
• Block write time
• Block read time
• Transactions committed / second
• Transactions rolled back / second
• Deadlocks / sec
• Cache hit ratio

Alerts

The following is a list of alerts available for PostgreSQL on Windows:

PostgreSQL specific alerts

• Cache hit ratio
• Configuration change
• Database unavailable
• Deadlock rate
• Error log entry
• Instance unreachable
• Internal Redgate Monitor error (instance)
• Limited sampling
• Long-running idle transaction
• Long-running query
• Long-running transaction
• Monitoring error (instance data collection)
• Monitoring stopped (instance credentials)
• Percentage of connections used
• Query throughput
• Replica not healthy
• Replica role change
• Replication lag
• Table bloat
• Transaction ID consumption

Host machine alerts

• Clock skew
• Disk average read time
• Disk average write time
• Disk space
• Host unreachable
• Internal Redgate Monitor error (host)
• Monitoring error (host data collection)
• Monitoring stopped (host credentials)
• Network utilization
• Physical memory
• Processor (CPU) under-utilization
• Processor (CPU) utilization