Metrics Monitoring on AWS and Ensuring Optimization
While Amazon’s services can afford to withstand higher activity and can adjust to meet increased throughput, it’s still advisable to watch activity carefully. Optimization of any applications built on this platform will lead to a lower monthly bill in the long term. Using AWS’ Relational Database Service (RDS) does come with its own set of metrics monitoring options to show users what resources they are proactively using.
This includes:
-
- CPU Utilization
- Queries
- Deadlocks
- Database Connections
- DML Throughput
- DDL Latency
- Available RAM
- Row Lock Time
- Network Throughput
- Active Transactions
CPU Utilization
As it’s plainly stated, this is the percentage of how much the CPU’s capacity is occupied by certain resources. Metrics are found here.
-
- Environment: Dev
- Cluster Identifier: oneriver-investor-portal-db-dev
- Alarm: High CPU Utilization threshold reached
- Criteria: Cpu utilization is >=25% in the last 5 mins
Queries
This metric counts the number of queries made per second. Metrics are found here.
-
- Environment: Dev
- Cluster Identifier: oneriver-investor-portal-db-dev
- Alarm: High number of Queries threshold reached
- Criteria: Number of queries >=10 in the last 5 mins
Deadlocks
Deadlock occurs when there are multiple database connections running queries on the same set of data. The first thread locks row X, then the second thread locks row Y, then the first thread tries to lock row Y and waits, then the second thread tries to lock row X and waits. Monitoring for such incidents is found here.
-
- Environment: Dev
- Cluster Identifier: oneriver-investor-portal-db-dev
- Alarm: Deadlocks threshold reached
- Criteria: Deadlocks >=2 in the last 5 mins
Database Connections
The number of connections to a Database instance. All connections are logged here.
-
- Environment: Dev
- Cluster Identifier: oneriver-investor-portal-db-dev
- Alarm: Max connections alarm threshold reached
- Criteria: Max number of db connections is >=2 in the last 5 mins
DML Throughput
This type of throughput covers the average number of times data is inserted, updated, and deleted per second. The average data interaction is logged here.
-
- Environment: Dev
- Cluster Identifier: oneriver-investor-portal-db-dev
- Alarm: Low DML Throughput Threshold reached
- Criteria: Avg number of DML <=0.5 in the last 5 mins
DDL Latency
Similar to DML Throughput, DDL Latency covers the average duration a data insert, update, or deletion constitutes. The average data interaction duration is logged here.
-
- Environment: Dev
- Cluster Identifier: oneriver-investor-portal-db-dev
- Alarm: High DDL Latency threshold reached
- Criteria: Avg DDL Latency >=25 in the last 5 mins
Freeable Memory
The amount of available random access memory. For MariaDB, MySQL, Oracle, and PostgreSQL DB instances, this metric reports the value of the MemAvailable field of /proc/meminfo. Available RAM is tracked here.
-
- Environment: Dev
- Cluster Identifier: oneriver-investor-portal-db-dev
- Alarm: Low Freeable Memory threshold reached
- Criteria: Freeable Memory <=500MB in the last 5 mins
Row Lock Time
{unable to find content on this category} Metrics are logged here.
-
- Environment: Dev
- Cluster Identifier: oneriver-investor-portal-db-dev
- Alarm: Row Locks threshold reached
- Criteria: Row Locks >=5 in the last 5 mins
Network Throughput
The amount of outgoing and inbound traffic to a specific database instance. This includes both customer database traffic and traffic associated with monitoring and replication. Traffic is monitored from here.
-
- Environment: Dev
- Cluster Identifier: oneriver-investor-portal-db-dev
- Alarm: Low Network Throuput threshold reached
- Criteria: Network Throughput is <=5 in the last 5 mins
Active Transactions
Counts the total active transactions. Everything is logged here.
-
- Environment: Dev
- Cluster Identifier: oneriver-investor-portal-db-dev
- Alarm: Active Transactions threshold reached
- Criteria: Active Transactions >=10 in the last 5 mins
Select Throughput
The average number of queries made per second. This metric is monitored from here.
-
- Environment: Dev
- Cluster Identifier: oneriver-investor-portal-db-dev
- Alarm: Select Throughput threshold reached
- Criteria: Select Throughput <=10 in the last 5 mins
