Monitoring AWS DynamoDB

As mentioned earlier, eG Enterprise provides a specialized model for monitoring AWS DynamoDB (see Figure 1).

Figure 1 : Layer model of AWS DynamoDB

Each layer of Figure 1 above is mapped to tests that report a variety of metrics revealing the availability, replication, resource usage and much more!

Using these metrics, administrators can find quick and accurate answers to the following persistent performance queries:

  • Are there too many tables with system error?

  • Are there too many user error and system errors on the target server?

  • Whether the network connection is available or not in the target server?

  • Does the database is taking too much time to respond to a query?

  • Whether the database connection is available or not?

  • Does the database is taking too much time to execute the query?

  • Are there too many throttled requests, conditional check failed requests, and transaction conflict requests on the target server?

  • Are there too many read and write throttled events?

  • Are the read and write capacity units rightly provisioned?

  • Are the number and percentage utilization of read and write capacity units too high? If so, then how much?

  • What is the rate of read and write capacity units consumption?

  • What is the percentage of completion when a new global secondary index is being added to a table?

  • Is the request latency too high?

  • Is the age of old unreplicated records too high? If so, then how much?

  • Whether or not replication enabled in the target server?

  • Are there too many unavailable regions on the target server? If so, then how many?

  • Are there too many pending replication by records on the target region?

  • Is the replication region available? What is the region response time?

  • Is the connection time to the replication region too high?

  • Does the region takes too much time to execute the query?

  • Is the replication latency on the target region too high?