We have some components with either a low traffic or predictable load. For some patterns, it makes more sense to go serverless, and for others to go semi-serverless and distribute load across servers using message queues. What considerations do we take into account to make a decision on this, and at what point does it make sense to switch from one to the other?
I’ll be talking about what role serverless plays in scaling iPlayer with an ever-changing feature set and varying audience patterns, and under which circumstances we have migrated away from serverless, as well as to serverless infrastructure.
To share best-practices around decision-making of when to use serverless, and share applicable patterns.
Basic knowledge of AWS technology is helpful but not strictly necessary.
You can view Nik’s slides below: