Study Guide · Bucket 7 of 7
Monitoring, Observability & Cost
Production mein sab kuch chalne ke baad bhi kaam khatam nahi hota — cheezein break hoti hain, bills shock deti hain, aur security audits chahiye hote hain. Yeh bucket 'operate' karna sikhata hai.
Context
Cloud = infrastructure APIs. Har action ek API call hai. Isliye monitor karna, cost track karna, aur audit karna — teeno equally zaroori hain jitna building karna.Amazon CloudWatchmonitoring & observability
What
AWS ka poora monitoring/observability hub. 4 pieces: Metrics (CPU, memory, disk performance), Logs (kya ho raha hai), Alarms (threshold cross hote hi notify), Dashboards (sab visualize karne ke liye).
Why
Problem ko user se pehle detect karna — yehi observability ka poora point hai. Bina monitoring ke production system blind flying hai.
How
Metrics automatically collect hote hain -> Alarm set karo (e.g. CPU > 80% for 5 min) -> Alarm SNS ke through notify karta hai (email/Slack) -> Dashboard pe sab visualize karo.
Skill
Concept-level, alarm configuration.
- Metrics vs Logs vs Alarms vs Dashboards
- SNS integration for alerts
Cost Explorer & Budgetscost management
What
AWS Budgets: custom spending budgets set karo, threshold cross hone pe alert. Cost Explorer: spending patterns visualize/forecast karo.
Why
AWS bills shock kar sakte hain — running resources bhool jaana ek common (aur mehenga) mistake hai. Interview mein bhi 'cost kaise control karoge' aksar poocha jaata hai.
How
Budgets mein ek monthly/custom threshold set karo -> jab spending us threshold ke paas aaye, alert milta hai -> Cost Explorer se dekho kaunsi service sabse zyada kharch kar rahi hai.
Skill
Concept-level — but practically enable zaroor karo apne free-tier account mein.
- Budget alarms
- Cost forecasting
Real org mein
Real org mein: FinOps ek separate practice hai. Engineers ko bhi cost-conscious hona expect kiya jaata hai — 'kya yeh resource zaroor chahiye' poochhna culture ka part hai, na ki sirf finance team ka kaam.AWS CloudTrailaudit & compliance
What
Har API call (har action jo tum AWS mein karte ho) ko record karta hai — governance, compliance, aur debugging ke liye.
Why
Agar koi production database accidentally delete kar de, CloudTrail se pata chal sakta hai kisne, kab, aur kaise. Security compliance ke liye bhi zaroori hai.
How
Ek baar enable karo (usually already on by default in modern accounts) — kuch specific kaam nahi karna, bas exist karta hai aur log karta rehta hai.
Skill
Awareness — enable karna aur logs padhna aana chahiye.
- Every action = an API call
- Immutable audit trail
Key Tasks — is bucket ko 'done' maanne se pehle
- CloudWatch alarm banao jo EC2 CPU > 70% pe notify kare
- AWS Budget set karo apne free-tier account pe (e.g. $5 threshold)
- CloudTrail event history mein apni recent activity dekho
- Ek dashboard banao jisme CPU + billing dono dikhein