Kashif manzer

@kashif_manzer

Software engineer across backend, distributed systems & AI/ML infra | Sharing what I learn building at scale Blockchain and Crypto enthusiasts.

Kashif manzer
@kashif_manzer
4h
47 microservices. 9 engineers. Checkout made 11 synchronous calls and email sat on the critical path. When email went down, checkout went down with it. How many of your services are secretly on the critical path?
2
Kashif manzer
@kashif_manzer
9h
One curl flag shows where your latency actually goes. DNS, connect, TLS, wait. No guessing. Which of these four is usually slowest for you?
Kashif manzer
@kashif_manzer
Nov 8
38 of 39 towns just voted to restrict data centres. The cloud is not abstract anymore. It is a building next door that drinks power and water. When did your town start arguing about servers?
Kashif manzer
@kashif_manzer
Nov 7
I am still learning how this all fits together, so seeing Actions and Pages slow down together on Nov 4 surprised me. What do you check first when a deploy stalls?
GitHub Status
@githubstatus
Nov 7
On November 4, 2026, between 15:49 UTC and 16:48 UTC, and again between 17:20 UTC and 17:55 UTC, GitHub Actions, Pages, Webhooks, Code Scanning, and Copilot coding agent experienced delays. During this time, Actions workflow runs and job status updates ... githubstatus.com/incidents/xr3m…
Kashif manzer
@kashif_manzer
Nov 6
Scaling from 12 to 30 pods made the outage worse. Three retrying layers meant up to 27 attempts per request. Autoscaling does not fix a retry storm. It feeds it. When did your autoscaler last make things worse?

Kashif manzer’s posts