I noticed that reading from Azure Append Blob is very slow when blob is added several thousand times or more. Writing / adding is quick, but reading a typical trinket for a magazine with several thousand adds each size to several KB, for a total size of several MB it takes more than a minute! Reading a standard or similar sized page blog takes milliseconds. Is there a way to speed up reading from append blob, i.e. by smoothing the internal structure?
So far, it seems that it is best to periodically “archive” / convert appob blobs to block blocks, and then process them. Any suggestion?
If not, what would an alternative log repository be recommended? You can use the Azure table, but more reading is required, even if bulk operations are used.
source
share