Mackerel supports obtaining and monitoring Amazon Relational Database Service (RDS) metrics in AWS Integration.
Please refer to the following page for AWS Integration configuration methods and a list of supported AWS services.
AWS Integration
Obtaining metrics
The metrics obtainable with AWS Integration’s RDS support are as follows. For Metric explanations, refer to the AWS help page.
| Graph name | Metric | Metric name in Mackerel | Unit | Statistics |
|---|---|---|---|---|
| CPU | CPUUtilization | rds.cpu.used | percentage | Average |
| CPU Credit | CPUCreditUsage CPUCreditBalance |
rds.cpu_credit.used rds.cpu_credit.balance |
integer | Average |
| Database Connections | DatabaseConnections | rds.database_connections.used | float | Average |
| Disk Queue | DiskQueueDepth | rds.disk_queue.depth | float | Average |
| BinLog Disk Usage | BinLogDiskUsage | rds.disk_usage.bin_log | bytes | Average |
| Memory | FreeableMemory SwapUsage |
rds.memory.free rds.memory.swap |
bytes | Average |
| Free Storage Space | FreeStorageSpace | rds.disk.free | bytes | Average |
| Replica Lag | ReplicaLag | rds.replica_lag.lag | float | Average |
| Disk IOPS | ReadIOPS WriteIOPS |
rds.diskiops.read rds.diskiops.write |
iops | Average |
| Disk Latency | ReadLatency WriteLatency |
rds.latency.read rds.latency.write |
float | Average |
| Disk Throughput | ReadThroughput WriteThroughput |
rds.throughput.read rds.throughput.write |
bytes/sec | Average |
| Network Throughput | NetworkReceiveThroughput NetworkTransmitThroughput |
rds.network_throughput.read rds.network_throughput.transmit |
bytes/sec | Average |
In addition to the above, metrics that can be obtained with Aurora are listed below.
| Graph name | Metric | Metric name in Mackerel | Unit | Statistics |
|---|---|---|---|---|
| Storage | FreeLocalStorage | rds.aurora.storage.free | bytes | Average |
| Volume Bytes Used | VolumeBytesUsed | rds.aurora.volume_bytes_used.used | bytes | Average |
| Volume IOPS | VolumeReadIOPS VolumeWriteIOPS |
rds.aurora.volume_iops.read rds.aurora.volume_iops.write |
iops | Average |
| BinLog Replica Lag | AuroraBinLogReplicaLag | rds.aurora.binlog_replica_lag.lag | float | Average |
| Replica Lag | AuroraReplicaLag AuroraReplicaLagMaximum AuroraReplicaLagMinimum |
rds.aurora.replica_lag.lag rds.aurora.replica_lag.max_lag rds.aurora.replica_lag.min_lag |
float | Average |
| Engine Uptime | EngineUptime | rds.aurora.engine_uptime.uptime | float | Average |
| Transactions | ActiveTransactions BlockedTransactions |
rds.aurora.transactions.active rds.aurora.transactions.blocked |
float | Average |
| Cache Hit Ratio | BufferCacheHitRatio ResultSetCacheHitRatio |
rds.aurora.hit_ratio.buffer rds.aurora.hit_ratio.result_set |
percentage | Average |
| DML Latency | InsertLatency UpdateLatency DeleteLatency DMLLatency |
rds.aurora.dml_latency.insert rds.aurora.dml_latency.update rds.aurora.dml_latency.delete rds.aurora.dml_latency.dml |
float | Average |
| DML throughput | InsertThroughput UpdateThroughput DeleteThroughput DMLThroughput |
rds.aurora.dml_throughput.insert rds.aurora.dml_throughput.update rds.aurora.dml_throughput.delete rds.aurora.dml_throughput.dml |
float | Average |
| Commit Latency | CommitLatency | rds.aurora.commit_latency.commit | float | Average |
| Commit Throughput | CommitThroughput | rds.aurora.commit_throughput.commit | float | Average |
| Select Latency | SelectLatency | rds.aurora.select_latency.select | float | Average |
| Select Throughput | SelectThroughput | rds.aurora.select_throughput.select | float | Average |
| DDL Latency | DDLLatency | rds.aurora.ddl_latency.ddl | float | Average |
| DDL Throughput | DDLThroughput | rds.aurora.ddl_throughput.ddl | float | Average |
| Queries | Queries | rds.aurora.queries.queries | float | Average |
| Login Failures | LoginFailures | rds.aurora.login_failures.failures | float | Average |
| Deadlocks | Deadlocks | rds.aurora.deadlocks.deadlocks | float | Average |
Precautions
Among the graphs/metrics obtainable with AWS Integration, metrics included in the following graphs are usually obtained in 5 minute intervals (only when using T2 instances).
- CPU Credit
- Volume IOPS
This is specified by the AWS CloudWatch API.
When creating monitoring rules for metrics with a 5-minute interval, the “condition duration” can only be configured for “monitoring an average value of 1 minute” (if any other value is selected, monitoring will not be done due to insufficient data points necessary for calculating the average value).
In addition, if this granularity is continually used, Mackerel metrics will be rounded to NULL as a result of the ratio of valid data points, and will not be displayed in graphs. (For more information on Mackerel’s metric retention granularity, please refer to the Mackerel Overview - Metric Retention)