/Tech1h ago

Coinbase Postmortem Reveals No Automated Failover In Single-Region Trading Engine

2018984621.7K
Original post
Gergely Orosz@GergelyOrosz#1443inTech

The postmortem from Coinbase's 10-hour outage is out and... damn

They run global trading from a single region because of latency. OK, I understand.

BUT they have no automated failover prepared!

Are they praying the region never goes down?? Doesn't compute for me...

5:44 AM · Jun 11, 2026 · 14.4K Views
Sentiment

Many users condemned Coinbase's single-region trading engine lacking automated failover as an unacceptable engineering failure for a large public company.

Pos
0.0%
Neg
100.0%
9 comments with sentiment.
Cluster Engagement
Posts from X
Most Activity
Most Activity
VIEWS7.4KBOOKMARKS8LIKES28REPLIES2
Gergely Orosz@GergelyOrosz

Full postmortem: https://www.coinbase.com/en-gb/blog/a-postmortem-of-our-may-7-2026-outage

I would get this for a small company.

But a $40B company? It's a serious WTH moment for me. And my impression of Coinbase engineering nosedived. This is resilliency basics: if you depend on a region, have auto failover and exercise it...

Gergely Orosz@GergelyOrosz

The postmortem from Coinbase's 10-hour outage is out and... damn

They run global trading from a single region because of latency. OK, I understand.

BUT they have no automated failover prepared!

Are they praying the region never goes down?? Doesn't compute for me...

1hViews 7.4KLikes 28Bookmarks 8
Gergely Orosz@GergelyOrosz

@escobyte unbelievable. this is such an engineering fail

1hViews 276Likes 2

@GergelyOrosz lol the same company which can't execute an order for 2 hours in a single region b/c of latency, ok.

Over the last 5 years, every usage attempt I had with Coinbase failed in some different way, at this stage I'm not really sure how they are still not busted.

47mViews 131Likes 2
Matija@mmunjako

@GergelyOrosz worked on a small startup project, from australia. they spent less than 1M on dev in total. having automatical failover across zones in the event of a complete AWS zone failure was necessary for investors to even want to look at it. and it wasn't even handling money. wtf coinbase

1hViews 164
Adnan M.@adeevicc

@GergelyOrosz I wonder how the damage from the outage compares to the benefit their customers and their business gained from the reduced latency.

58mViews 101
VLAD ARBATOV@vladzima

@GergelyOrosz single-region matching is normal in trading, nasdaq and nyse also run one primary site since an order book is a sequence. but no warm standby with sequenced replication... exchanges have done that for decades

1hViews 58
frustasean@Frustasean

@GergelyOrosz Use CPU or GPU to compute

55mViews 46
Johnny Utah@johnnyutah269

@GergelyOrosz Ehhhhh compute lol nice dad joke

51mViews 41
Esco Obong@escobyte

@GergelyOrosz No automate failover! That had to be it. I discussed this when it happened

1hViews 31
Joaquín Corredor@grexcorvorum

@GergelyOrosz I don't understand how that passes SOX IT controls for a public company.

41mViews 30
svgmzr@svgmzr

@GergelyOrosz That's common thread in modern tech. How a guy can write a working game in his basement yet Facebook can't make their primitive app work properly?

1hViews 19
Atul Khola 💊@pixelandpump

@GergelyOrosz resource "aws_faith" "us_east_1" {}

1hViews 19
Bitcoin Is Property@CivilBtcSaver

@GergelyOrosz stfu these things are hard

don’t pretend you would have done better

39mViews 9
Shane@wherethefubeen

@GergelyOrosz Even German mid-market companies do multi-cloud because of the “what if AWS goes down tomorrow?” scenario, and a financial institution of their size is thinking only about the happy path?

1hViews 3
essa@Mrojek1

@GergelyOrosz Recently they’ve said that non-technical people write code for them as they laid a bunch, what did you expect? 😆

1hViews 2
mellanox@mellanoxx

@GergelyOrosz Last night I was studying on fault tolerance. Didn't know would see an example right next day

1h