General Windows Azure Best Practices :
-> Understand Account Scalability targets.
- Use multiple storage accounts to get more
- Distribute your storage accounts across the region.
-> Cache critical data sets
- As a "backup" data set to fall back on
- To get more request/sec than the account/partition target
-> Distribute load over many partitions and avoid spikes
- Avoid append/Pretend patterns
-> Locate storage accounts close to compute/users
-> Use HTTPS
-> Optimize what you send and receive
-Blobs: Range reads, Metadata, Head Requests
-Tables: JSON vs AtomPub,Upsert,Merge,Projection<Point Queries
-Queses: Update Message,Batch size
->Control Parallelism at the application layer
-Unbounded Parallelism can lead to slow latency and throttleing
-> Enable logging & Metrics
- Can be done via REST, Client API or Portal
- Enables clients to self diagnose issues, including performance related ones
- Data can be automatically GC'd according to a user specified retention interval
- Have longer retention for hourly metrics and shorter retention for minute metrics.
-> To build resilient distributed system
-Rely on caching
* Caching layer with fail fast but cache updated in background
-> Use storage Analysts to learn about your usage
-understand limits of single storage account , partition
-Avoid unnecessary round trips & bandwidth usage
-continuously monitor for anomalies
-> Efficient retries
-> Exponential back-off for background job
-Constant back-off for websites
-Use secondary for higher availability
Hope this helps!
Regards.
