Black Friday cloud scaling

I Googled for a relevant image and found this gen-AI-created one from Ops Digest, which cracked me up – but the accompanying blog post is 100% worth a read! After mine, of course :)

Friendly reminder to everyone who has important stuff deployed in the cloud – Black Friday/Cyber Monday is coming! Do you have any reservations in place that guarantee sufficient compute capability for your mission-critical IaaS apps? Do you have auto-scaling configured (both metrics-based and scheduled per historical traffic) enough to handle a potentially unexpected spike in demand?

Even if you’re not in e-commerce, sharing public cloud resources with other customers who might be selling stuff around Black Friday means you should stake your claim to make sure your apps of any kind keep working while everyone frantically buys a discounted air fryer. If you PaaS folks are feeling smug about not having to deal with configuring scaling, please at least consider double-checking any quotas or limits that would prevent the burst you need in a pinch. :) And if you use GKE, check out this blog post and the related dashboard and report; if you’re running k8s elsewhere, the high-level info is still useful for prioritising important pods.

None of this is meant to panic you about the cloud – after all, this would all be SO much more annoying and unreliable and challenging to deal with on premises (my sympathies to you datacenter elves). But it’s a good idea not to rest on your laurels, especially if you’ve migrated a workload to cloud recently enough that you haven’t yet dealt with a Black Friday cycle. A couple quick checks will make you (and your staff) feel calmer as the season progresses.

♡ your friendly holiday cloud fairy; apologies from America for the consumerism-induced resource panic

This post appeared initially on LinkedIn; I finally migrated it here.

Leave a comment

Your email address will not be published. Required fields are marked *