Uber migrates microservices to multi-cloud platform running Kubernetes and Mesos

Post by leesalminen » Mon, Oct 23, 2023, 3:46 AM UTC

charcircuit wrote: post #195

scarface_74 wrote: post #190

Earlier quoted context omitted.

You’ve never been the one neck to choke when things go wrong have you? If Billy Bob’s cloud provider goes down, you are going to constantly be blamed for making a poor decision. If anything goes wrong they are going to question your decision. If you choose AWS (or Azure) and a region goes down - everyone else is down too. “No one ever got fired for choosing IBM”. Choosing the most popular vendor - AWS, Salesforce, Se…

Even if the alternate cloud provider goes offline for an entire day it still would be worth it financially compared to AWS because egress is so expensive there.

Post by insanitybit » Mon, Oct 23, 2023, 3:56 AM UTC

charcircuit wrote: post #186

scarface_74 wrote: post #181

Earlier quoted context omitted.

And when has that happened with respect to either GCP, AWS or Azure at a level that it’s worth migrating? Even if you have done everything in a “cloud agnostic” way, “infrastructure has weight”. Any large migration isn’t just technical , it involves project management, organization training, regression testing, compliance testing, security testing, architecture review boards, vendor negotiations, firewall changes, co…

>And when has that happened with respect to either GCP, AWS or Azure at a level that it’s worth migrating? Egress price makes it worth migrating away from those three.

Post by thakoppno » Mon, Oct 23, 2023, 4:31 AM UTC

oooyay wrote: post #95

x86x87 wrote: post #40

Earlier quoted context omitted.

Question for you: how does performance measurement and optimization work in that environment? Is the key some sort of meta tooling that understands relationships between microservices? How would you express such relationships in the first place? Q2: How do you ensure the stubbed deps behave like the real thing? Q3: how do you handle logging and metrics in an unified way across the stack? And related to this: how do y…

> how does performance measurement and optimization work in that environment? SRE here. Generally speaking, each API or each service will have a contract that it must adhere to depending on upstream and downstream relationships and their fail safes. Each service (or API) will then load test in isolation. After that, if you want to be really sure about regressions (which would include fail safes) you load test the who…

Post by scarface_74 » Mon, Oct 23, 2023, 4:53 AM UTC

insanitybit wrote: post #213

charcircuit wrote: post #186

Earlier quoted context omitted.

>And when has that happened with respect to either GCP, AWS or Azure at a level that it’s worth migrating? Egress price makes it worth migrating away from those three.

Egress prices make migrating away hard. But a lot of products don't need to push much data out of AWS, especially with VPC peering based products.

Post by isbvhodnvemrwvn » Mon, Oct 23, 2023, 6:21 AM UTC

lamontcg wrote: post #212

pc86 wrote: post #165

Earlier quoted context omitted.

Of course you could, just like you could do this in 2000 less-well-defined microservices, or 8000 more finely-grained ones. The question is what makes you think 1 service is immediately better than however many payment services there are now?

All other things being equal, 1 service is obviously better than 4,000 services to maintain.