Recently a few customer have started rolling out agileBase for high volume processes, which means we see charts of activity (logins, data views, data edits, report building etc.) like these

Company 1

logins

and these

Company 2

logins 2

In total, for an average day, the number of data edits per hour look something like

edits per hour

i.e. a sudden rise in the morning between 7am and 9am, followed by a larger just after lunch.

On our side monitoring the server, we see key metrics like

average no. database queries per minute = 500

with some sharp spikes (caused mainly by schema changes rather than user activity) of over 10000/min.

sql hits per hour

SQL queries per minute 

Luckily, the average time a query takes to execute is just 10ms, that’s 100th of a second.

sql mean times

Average SQL query duration (ms)

Nonetheless, to keep things running smoothly and to make sure we can deal with near-future growth, the memory capacity of our cloud service has been doubled and the processing power boosted by an extra 50%.

Advertisements