Harshita
SDET | Playwright, Selenium, REST Assured | Quality is a team sport
Caching dependencies properly, running tests in parallel shards and only building Docker images on main. Devs are noticeably happier.
Kubernetes, Terraform, AWS, CI/CD and on-call survival
33 members 8 postsTechnology
Practical DevOps and SRE discussions: Kubernetes, cloud cost, Terraform, observability, incident reviews. Share your setups and postmortems (sanitised please).
Harshita
SDET | Playwright, Selenium, REST Assured | Quality is a team sport
Caching dependencies properly, running tests in parallel shards and only building Docker images on main. Devs are noticeably happier.
Aditya Nair
DevOps / SRE | Kubernetes, Terraform, AWS | Making deploys boring since 2019
What changed our on-call from dreaded to fine: • Every alert must be actionable, or it gets deleted • Runbook link in every alert • Weekly review of all pages • Comp day after a bad night
Siddharth Saxena
AppSec Engineer @ Moneyleaf | OWASP, threat modeling, bug bounty | Making security developer-friendly
Audited our CI last week: 4 long-lived cloud keys, one 3 years old, one belonging to someone who left in 2024. Moved everything to OIDC federation (GitHub Actions → AWS role). No static keys left. Took one afternoon.
Karan Bhatia
Cloud Architect | AWS, Azure, Security | Helping enterprises move to cloud without lighting money on fire
Starting a new landing zone for a client. They're wary of licence changes. Anyone running OpenTofu in production yet?
Karan Bhatia
Cloud Architect | AWS, Azure, Security | Helping enterprises move to cloud without lighting money on fire
1. Root account has no MFA or has access keys 2. S3 buckets with public ACLs 'just for testing' 3. Security groups with 0.0.0.0/0 on 22 or 5432 4. No CloudTrail in all regions 5. IAM users with AdministratorAccess for CI 6. No budget alerts
Harshita
SDET | Playwright, Selenium, REST Assured | Quality is a team sport
Our Playwright suite gates prod deploys. 1-2% flakiness means a couple of blocked deploys a week. Devs want to auto-retry, I'm worried retries will hide real bugs. How do others handle this?
Aditya Nair
DevOps / SRE | Kubernetes, Terraform, AWS | Making deploys boring since 2019
Six things, ranked by impact: 1. Right-sized requests using 2 weeks of actual usage (most pods requested 3-4x what they used) 2. Moved stateless workloads to Spot via Karpenter with diverse instance types 3. Scaled dev/staging to zero outside 9am-9pm IST 4. Switched to Graviton (arm64) for 70% of services, just a
Aditya Nair
DevOps / SRE | Kubernetes, Terraform, AWS | Making deploys boring since 2019
Kubernetes, Terraform, CI/CD, observability, cost and on-call. Share your setups, ask questions, post (sanitised) postmortems. First question for everyone: what's the one alert in your setup that pages way too often?