Pekik Indra Lesmana @_indrax - Twitter Profile

we all know that AWS went down for some time. let’s look at what actually happened: The US-EAST-1 region is one of AWS’s largest and most critical data centers, hosting infrastructure for thousands of major websites and applications. When DynamoDB, a key database service, began experiencing high error rates, it created a ripple effect across AWS’s interconnected systems. This caused widespread disruptions because many applications rely on DynamoDB for data storage and retrieval operations. You see, DynamoDB is deeply embedded in AWS’s control plane architecture. Many AWS services use it internally for storing metadata, state information, and service configurations. So, when the DynamoDB API endpoint became unreachable via DNS, these dependent services couldn’t access critical operational data. The failure affected over 36 AWS services simultaneously. When DNS resolution failed, client services couldn’t locate the DynamoDB endpoints. This triggered a feedback loop similar to the September 2015 DynamoDB outage. Each retry consumed more resources, causing more timeouts, which in turn triggered further retries, creating a vicious cycle. AWS responded by pursuing multiple parallel recovery paths simultaneously, rather than waiting to identify and fix a single root cause. The primary fix involved stabilizing the DNS resolution infrastructure for the DynamoDB endpoint. AWS’s engineering teams likely: -Reconfigured or restarted DNS resolvers handling DynamoDB. -Potentially rerouted DNS queries through alternative resolution paths. -Implemented emergency DNS caching adjustments. This incident reflects the interdependence and fragility of large-scale cloud ecosystems, where even a single point of failure in a core service can cascade across dozens of systems within minutes.

codoyevskyy's tweet photo. we all know that AWS went down for some time.

let’s look at what actually happened:

The US-EAST-1 region is one of AWS’s largest and most critical data centers, hosting infrastructure for thousands of major websites and applications.

When DynamoDB, a key database service, began experiencing high error rates, it created a ripple effect across AWS’s interconnected systems. This caused widespread disruptions because many applications rely on DynamoDB for data storage and retrieval operations.

You see, DynamoDB is deeply embedded in AWS’s control plane architecture. Many AWS services use it internally for storing metadata, state information, and service configurations. So, when the DynamoDB API endpoint became unreachable via DNS, these dependent services couldn’t access critical operational data. The failure affected over 36 AWS services simultaneously.

When DNS resolution failed, client services couldn’t locate the DynamoDB endpoints. This triggered a feedback loop similar to the September 2015 DynamoDB outage. Each retry consumed more resources, causing more timeouts, which in turn triggered further retries, creating a vicious cycle.

AWS responded by pursuing multiple parallel recovery paths simultaneously, rather than waiting to identify and fix a single root cause.

The primary fix involved stabilizing the DNS resolution infrastructure for the DynamoDB endpoint. AWS’s engineering teams likely:
-Reconfigured or restarted DNS resolvers handling DynamoDB.
-Potentially rerouted DNS queries through alternative resolution paths.
-Implemented emergency DNS caching adjustments.

This incident reflects the interdependence and fragility of large-scale cloud ecosystems, where even a single point of failure in a core service can cascade across dozens of systems within minutes.

65

4K

625

2K

469K

_indrax retweeted

Sammy Notaslimboy

@NOTASLIMBOY

about 1 year ago

Ini kronologis sampai bisa jadi seperti hari ini. Makanya jangan jumawa dan flexing aja tuh kalian yang sedang di atas. 1.Katanya efisiensi, tapi jumlah menteri banyak, wamen-wamen dikasih jabatan komisaris BUMN. 2. Kenaikan PBB di berbagai daerah, sampai nilai yang fantastis, memicu demo besar terutama di Pati. 3. Di tengah himpitan ekonomi, harus menyaksikan kenaikan tunjangan DPR yg tidak masuk akal. Eh, anggota dewannya pada joget lagi di Senayan. Setelahnya banyak yang ngeles bela diri dengan statemen nggak masuk akal. 4. Provokasi bbrp anggota DPR termasuk Sahroni. 5. Pecah demo besar, jatuh korban jiwa yang dilindas Rantis Brimob. 6. Demo makin besar.

86

6K

5K

386

207K