Infrastructure

Not all load balancing algorithms are equal. The right choice depends on your traffic patterns, backend capabilities, and consistency requirements. Round Robin The default. Requests go to each server in turn. R R R R e e e e q q q q u u u u e e e e s s s s t t t t 1 2 3 4 → → → → S S S S e e e e r r r r v v v v e e e e r r r r A B C A Nginx: ...

Deploying shouldn’t mean downtime. Blue-green deployment lets you release new versions instantly and roll back just as fast. The Concept You maintain two identical production environments: Blue: Currently serving live traffic Green: Idle, ready for the next version To deploy: Deploy new version to Green Test Green thoroughly Switch traffic from Blue to Green Green is now live; Blue becomes idle Next deploy: repeat with roles reversed B A e f f U t U o s e s r e r e e r r s s s d w e → i → p t l L c L o o h o y a : a : d d B B a a l l a a n n c c e e r r → → [ [ [ [ B G B G l r l r u e u e e e e e n n v ] v 1 1 v . i . 1 0 d 0 . ] l ] 1 e ] ✓ i d ✓ L l I e L V I E V E Implementation with Nginx Simple traffic switching with upstream blocks: ...

We’ve all done it. That first deployment where the database password lives in a .env file. The API key hardcoded “just for testing.” The SSH key committed to the repo because you were moving fast. Environment variables as secrets storage is the gateway drug of bad security practices. Let’s talk about what actually works. The Problem with Environment Variables Environment variables seem safe. They’re not in the code, right? But consider: ...

Terraform state is the source of truth for your infrastructure. Mess it up and you’ll be manually reconciling resources at 2 AM. Here’s how to manage state properly from day one. What Is State? Terraform state maps your configuration to real resources: 1 2 3 4 5 # main.tf resource "aws_instance" "web" { ami = "ami-12345" instance_type = "t3.micro" } 1 2 3 4 5 6 7 8 9 10 11 12 13 14 // terraform.tfstate (simplified) { "resources": [{ "type": "aws_instance", "name": "web", "instances": [{ "attributes": { "id": "i-0abc123def456", "ami": "ami-12345", "instance_type": "t3.micro" } }] }] } Without state, Terraform doesn’t know aws_instance.web corresponds to i-0abc123def456. It would try to create a new instance every time. ...

Every modern service needs health check endpoints. Load balancers probe them. Kubernetes uses them. Monitoring systems scrape them. But a naive implementation—returning 200 OK if the process is running—tells you almost nothing useful. Here’s how to build health checks that actually help. Two Types of Health Liveness: Is the process alive and not deadlocked? Readiness: Can this instance handle requests right now? These are different questions with different answers: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 # Liveness: Am I alive? @app.get("/health/live") def liveness(): # If this returns, the process is alive return {"status": "alive"} # Readiness: Can I serve traffic? @app.get("/health/ready") def readiness(): checks = { "database": check_database(), "cache": check_cache(), "disk_space": check_disk_space(), } all_healthy = all(c["healthy"] for c in checks.values()) return JSONResponse( status_code=200 if all_healthy else 503, content={"status": "ready" if all_healthy else "not_ready", "checks": checks} ) Why separate them? ...

Opening a database connection is expensive. TCP handshake, SSL negotiation, authentication, session setup—it all adds up. Do that for every query and your application crawls. Connection pooling fixes this by reusing connections. Here’s how to do it right. The Problem Without pooling, every request opens a new connection: 1 2 3 4 5 6 7 8 # BAD: New connection per request def get_user(user_id): conn = psycopg2.connect(DATABASE_URL) # ~50-100ms cursor = conn.cursor() cursor.execute("SELECT * FROM users WHERE id = %s", (user_id,)) user = cursor.fetchone() conn.close() return user At 100 requests per second, that’s 100 connections opening and closing per second. Your database server has a connection limit (typically 100-500). You’ll exhaust it fast. ...

Configuration is where deployments go to die. A typo in an environment variable, a missing secret, a config file that works in staging but breaks in production. Here’s how to make configuration boring and reliable. The Hierarchy of Configuration Not all config is created equal. Layer it: 1 2 3 4 5 . . . . . D C E C F e o n o e f n v m a a f i m t u i r a u l g o n r t n d e s f m - i e l f ( l n i l i e t n a n s e g v s c ( a f o p r l ( d e i a r e r a g u ) b s n e l t n e i v s m i e r ) o n m e n t ) S O D a R n y f E u e n e n n - a s v t o m t i i f i r m f c f o e a n c l m v h l e v e a b n e r n a t r r g c - r i e k s i d s s p d e e e s c s i f i c Later layers override earlier ones. This lets you: ...

Rate limiting is the bouncer at your API’s door. Too strict, and legitimate users get frustrated. Too loose, and one bad actor can take down your service. Here’s how to find the balance. Why Rate Limit? Without limits, a single client can: Exhaust your database connections Burn through your third-party API quotas Inflate your cloud bill Deny service to everyone else Rate limiting isn’t about being restrictive—it’s about being fair. ...

Every production system eventually needs background jobs. Email notifications, report generation, data syncing, webhook processing—the work that can’t (or shouldn’t) happen during a user request. Here’s what I’ve learned about making them reliable. The Naive Approach (And Why It Breaks) Most developers start with something like this: 1 2 3 4 5 @app.route('/signup') def signup(): user = create_user(request.form) send_welcome_email(user) # Blocks the response return redirect('/dashboard') This works until it doesn’t. The email service has a 5-second timeout. Now your signup page feels broken. Or the email service is down, and signups fail entirely. ...

DNS issues are deceptively simple. “It’s always DNS” is a meme because it’s true. Here’s how to actually debug it. The Essential Commands dig: Your Primary Tool 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 # Basic lookup dig example.com # Short answer only dig +short example.com # Specific record type dig example.com MX dig example.com TXT dig example.com CNAME # Query specific nameserver dig @8.8.8.8 example.com # Trace the full resolution path dig +trace example.com Understanding dig Output ; ; ; ; e ; ; ; ; ; e ; x ; ; ; ; x a Q a A m Q S W M U m N p u E H S E p S l e R E G S l W e r V N D T e E . y E : S i I . R c R I G O c o t : S Z N o S m i a E 9 m E . m 1 t . S . C e 9 1 E T : 2 F r 8 C I . e c . T O 2 1 b v 1 I N 3 6 d O : 8 2 : N m . 8 : s 1 5 e . 2 6 c 1 0 # : e 5 3 x 3 3 0 a 6 ( : m 0 1 0 p 0 9 0 l 2 e . E . 1 S c 6 T o I I 8 m N N . 2 1 0 . 2 1 6 ) A A 9 3 . 1 8 4 . 2 1 6 . 3 4 Key fields: ...