Having your own auth workflow was instant fail with the well architected framework committee. Using Okta was instant pass.
I don't necessarily disagree with that policy but given that Okta was breached several times while I was working there, it was interesting the extent to which our CSO had blinders about it.
Liability is the answer! If you build an auth system and it fails, it's your backside. If Okta fails, it's theirs. Enterprises buy products as much as they buy protection from problems.
e.g. when Crowdstrike takes down Windows across the worlds or AWS east coast falls over everybody hurts. At that point the story is easy, you point at the broken thing, mumble something about improving resilience, and everyone just moves on.
Roll your own system and have it taken down / breached specifically? There's noone to point at. It's hard to make the narrative anything except it being your fault.
You have (the perception of having) someone to forward the claim to once you're hit by one where the damages are quantified in money like a life insurance or disability payout caused by the data loss?
As someone that is not really in the game, does Okta have such a bad track record, and are there alternatives that are considered solid? From the outside, it seemed like EntraID is a bit of a burning dumpster fire, while Okta seemed expensive, but usable and decent (from comments I read)
The current default for lazy enterprise customers seems to be an unholy tangle of Active Directory, Entra, and Okta. If you use all three it's 3x more secure, right?
Having your own auth workflow was instant fail with the well architected framework committee. Using Okta was instant pass.
I don't necessarily disagree with that policy but given that Okta was breached several times while I was working there, it was interesting the extent to which our CSO had blinders about it.