3 ms·

I don't think blaming AWS is fair, since they typically exceed their regional and AZ SLAs AWS makes their SLAs & uptime rates very clear, along with explicit w

I don't think blaming AWS is fair, since they typically exceed their regional and AZ SLAs AWS makes their SLAs & uptime rates very clear, along with explicit warnings about building failover / business continuity. Most of the questions on the AWS CSA exam are related to resiliency . Look, we've all gone the lazy route and done this before. As usual, the problem exists between the keyboard and the chair.

dijit 1y ago
Not sure any of their SLA’s are covered here. If they don’t obfuscate the downtime (they will, of course), this outage would put them at, what, two nines? Thats very much out of their SLA. People also keep talking about it as if its one region, but there are reports in this thread of internal dependencies inside AWS which are affecting unrelated regions with various services. (r53 updates for example)
- tonymet 1y ago
  Sounds like your lesson is "yes we should continue shaming AWS rather than fix our app"
  - rester324 1y ago
    It sounds like you think the SLA is just toilet paper? When in reality it's a contract which defines AWS's obligations. So the lesson here is that they broke their contract big time. So yes. Shaming is the right approach. Also it seems you missed somehow the other 1700+ comments agreeing with shaming
    - tonymet 1y ago
      I wouldn't go that far. The SLA is a contract, and they are clear on the remedy (up to 100% refund if they don't hit 95% uptime in a month). Just like reading medication side effects, they are letting you know that downtime is possible, albeit unlikely. All of the documentation and training programs explain the consequence of single-region deployments. The outage was a mistake. Let's hope it doesn't indicate a trend. I'm not defending AWS. I'm trying to help people translate the incident into a real lesson about how to proceed. You don't have control over the outage, but you do have control over how your app is built to respond to a similar outage in the future.
Tractor8626 1y ago
They supposed to exceed their SLA. SLA is guaranteed worst case.
- tonymet 1y ago
  If you don't break anything, you aren't making anything valuable.

dijit 1y ago

Not sure any of their SLA’s are covered here. If they don’t obfuscate the downtime (they will, of course), this outage would put them at, what, two nines? Thats very much out of their SLA. People also keep talking about it as if its one region, but there are reports in this thread of internal dependencies inside AWS which are affecting unrelated regions with various services. (r53 updates for example)

tonymet 1y ago
Sounds like your lesson is "yes we should continue shaming AWS rather than fix our app"
- rester324 1y ago
  It sounds like you think the SLA is just toilet paper? When in reality it's a contract which defines AWS's obligations. So the lesson here is that they broke their contract big time. So yes. Shaming is the right approach. Also it seems you missed somehow the other 1700+ comments agreeing with shaming
  - tonymet 1y ago
    I wouldn't go that far. The SLA is a contract, and they are clear on the remedy (up to 100% refund if they don't hit 95% uptime in a month). Just like reading medication side effects, they are letting you know that downtime is possible, albeit unlikely. All of the documentation and training programs explain the consequence of single-region deployments. The outage was a mistake. Let's hope it doesn't indicate a trend. I'm not defending AWS. I'm trying to help people translate the incident into a real lesson about how to proceed. You don't have control over the outage, but you do have control over how your app is built to respond to a similar outage in the future.

Tractor8626 1y ago

They supposed to exceed their SLA. SLA is guaranteed worst case.

tonymet 1y ago
If you don't break anything, you aren't making anything valuable.