Effects
Increased latency can cause performance issues in your application. If your application relies on blob storage for unmanaged disks, then disk access will be impacted. Any blob, file, queue, or table operation latency can cause cascading slowdowns in your application.
Fixes
The Success E2E Latency metric measures the total amount of time it takes for requests to be processed by the storage account APIs, sent to the client, and then acknowledged by the client. This metric should usually track closely to the Average Success Server Latency metric, and the difference between the two metrics is usually the only area you can optimize.
Since end-to-end latency includes time spent transmitting data over the network, large requests will typically have much higher latency. High latency can also be caused by hitting the Storage Account limits. If you are using Standard performance, you can try upgrading to Premium performance to reduce latency. To achieve higher throughput with storage accounts past the limits, you can either open up a support request to ask for higher limits or shard your data between multiple storage accounts to get around the limits.
If you want to optimize for lower end-to-end latency, you can achieve significant results by tuning your client to make smaller requests for blobs, files, queues, and tables. Your success in this area is mostly dependent on your ability to change your application and what kind of data you are actually storing and accessing in storage accounts.
If access patterns have not changed for your application, but latency is unusually high, check the Azure Status page for any incidents related to Storage Accounts. If service is only impacted in one region, and you have the ability to failover your Storage Account to another region, consider going through the failover process so your requests go to the secondary location until the issue is resolved. Open up a support ticket with Azure Support to document the issue and get it resolved ASAP.
Resources
