The following is a summary of rules and advisories available out-of-the-box with the PostgreSQL cartridge. Default threshold values can be changed or scoped to specific values, generally through registry variables. These rules can be copied, modified, disabled, or customized in a wide variety of ways. For more information, refer to Working with Alarms.
This section describes the following rules and advisories:
Alerts if the connection count is approaching capacity.
Alerts if the backend write percentage is too high.
Alerts if the average growth of the tablespaces will cause the database server to fill up in the near term.
Alert if the average wait time for a statement is significantly longer than its historical average.
Alerts if the percentage of queries in a waiting state is too high.
Alerts if the total index size for a table is too large compared to the total table size.
Alerts if logical replication slot has become conflicted.
Alerts if the buffer hit percentage has been too low recently.
Alerts if the buffer hit percentage has been too low recently.
Alerts if a low percentage of checkpoints that are being performed are actually required.
Alerts if the agent has not been able to connect to the PostgreSQL database server.
Alerts if there are any ungranted locks older than the deadlock timeout setting.
Alerts if the difference in WAL segments between your primary’s sent_location and standby server’s replay_location is too great.
Alerts if safe_wal_size is approaching 0, assuming max_slot_wal_keep_size is not unlimited.
Alerts with warning if replication slot count is at max_replication_slots or less than PG_SlotCountAvailableWarn percentage, whichever is greater.
Alerts if a significant percentage of transaction data for a logical replication slot is exceeding available memory.
Alerts if WAL Status of replication slot is not in reserved state.
Alerts if it has been taking too long to establish a connection to the PostgreSQL server.
Checks for tables that may be experiencing auto analyze issues.
Checks for adequate memory allocation to autovacuum process based on largest table size.
Checks for tables that may be experiencing autovacuuming issues.
Checks for a high percentage of buffer writes by backends.
Checks for average connection usage close to max_connections.
Shows 10 most frequently executed queries in the last week.
Checks for large tables that would possibly benefit from creating indexes.
Shows 10 slowest queries by average time executed in the last week.
Checks for high number of pending I/O requests.
Checks for queries writing to temporary disk files.