Monitoring the Mongo on Azure Cosmos

eG Enterprise offers a special-purpose monitoring model for the Mongo Azure Cosmos Service with the help of which administrators can keep a constant check on the availability and overall performance of the service.

Figure 1 depicts the layer model of a Mongo Azure Cosmos Service

Figure 1 : The layer model of a Mongo Azure Cosmos Component

Each layer of Figure 1 above is mapped to tests that provide deep insights into how the Mongo Azure Cosmos Service manages connections to it, number of databases created, updated, deleted, throughput, number of database collections created, updated, deleted, number of connections available, number of records, database service availability and in the process provides effective pointers to how the service can be fine-tuned to perform better.

The metrics reported by these tests help answer the following performance questions:

• Is the storage size nearing capacity limits?

• Is the provisioned RU/s being fully consumed or exceeded?

• Are any requests being throttled due to RU limits (e.g., receiving 429 errors)?

• Are read or write operation latencies exceeding acceptable thresholds?

• Is any region or collection currently unavailable?

• Is there replication lag or failure between regions?

• Are there a high number of failed requests (e.g., 4xx or 5xx errors)?

• Are any queries running without using indexes, leading to performance degradation?

• Have there been any unauthorized access attempts or suspicious login activities?

• Have recent backups failed or not been validated for successful restoration?