LinkedIn Carousel: SLA Breach Lifecycle

Post type: Technical · 10 slides
Blog post: https://www.nuvikatech.com/blog/posts/sla-breach-lifecycle

SLIDE 1 — Cover

Headline: Detecting an SLA breach is the easy part. Here’s everything that comes after.

Sub-line: The full lifecycle from cloud incident to credit claim — and why most enterprises lose the money at step 2.

SLIDE 2

Label: THE FOUR STEPS

Headline: Detect → Quantify → File → Poll. Each step has hidden complexity.

Body: Detection is straightforward. The complexity lives in quantification (cross-month breaches, credit tiers), filing (per-provider APIs, deadline windows), and polling (provider vocabulary normalisation).

SLIDE 3

Label: STEP 1: DETECT

Headline: When uptime drops below the SLA threshold, open a breach record.

Body: 99.9% for most AWS services. 99.95% for certain Azure tiers. The breach stays open until the service recovers, then closes automatically.

SLIDE 4

Label: STEP 2: QUANTIFY

Headline: Cross-month breaches are the complexity nobody talks about.

Body: If a breach starts January 28th and ends February 3rd, the credit uses January’s spend for January’s downtime and February’s spend for February’s. We split the record at the month boundary automatically.

SLIDE 5

Label: CREDIT TIERS

Headline: AWS EC2: 10% credit at 99.1% uptime. 100% credit below 95%.

Body: Each provider publishes credit tiers. We maintain these in a registry and calculate against actual billing data. A 10% credit on a $50k/month EC2 spend is $5,000. That’s real money.

SLIDE 6

Label: STEP 3: FILE

Headline: GCP’s filing window is 30 days. Miss it and the credit is forfeited — no exceptions.

Body: AWS and Azure allow 60 days. GCP’s 30-day window is the dangerous one. Our system checks the window before every filing attempt and fires reminder alerts as the deadline approaches.

SLIDE 7

Label: STEP 4: POLL

Headline: Providers don’t send webhooks for claim approvals. You have to poll.

Body: A Cloud Scheduler job fires every day, polls all “Filed” breaches, and normalises each provider’s status vocabulary to approved / denied / pending / unknown.

SLIDE 8

Label: THE STATE MACHINE

Headline: Every path leads to a defined, actionable state. No stuck records.

Body: Pending → Active → Resolved → Filed → Credit Approved. Alternate paths: Assisted Filing Required (API failure), Expired (missed window), Credit Denied. Each triggers the appropriate alert.

SLIDE 9

Label: WHY IT MATTERS

Headline: Most enterprises lose SLA credits because nobody owns the process.

Body: The credits exist. The provider owes them. But monitoring three vendor support portals across time-sensitive windows is easy to deprioritise. Automation is the only reliable fix.

SLIDE 10 — CTA

Headline: Want the full story?

Body: The full lifecycle breakdown: detection, month-boundary splits, credit calculation, per-provider filing, and the polling architecture that tracks approvals.

Link: nuvikatech.com/blog/posts/sla-breach-lifecycle