Increased Queue API Latency After Migrating Storage Account from GRS to RA-GRS
Hello Storage Account team,
We recently migrated a storage account (Account A) to a new storage account (Account B). After the migration, we observed higher latency, particularly for queue-related APIs (e.g., PutMessage / GetMessage), on Account B.
So far, the only configuration difference we have identified between the two accounts is the replication setting:
- Account A: Geo-redundant storage (GRS)
Account B: Read-access geo-redundant storage (RA-GRS)
According to the documentation, write operations should be served from the primary region for both GRS and RA-GRS, and therefore no additional write latency is expected between the two replication options.
Given this, we would like to ask:
Could RA-GRS have any impact on Queue service write latency in practice?
Are there any known internal behaviors, throttling, replication side effects, or other factors we should investigate when seeing increased Queue latency after such a migration?
Any guidance or pointers on where to continue troubleshooting would be greatly appreciated.
Thank you for your help.