Zhivko Todorov
ALL CASE STUDIES

CASE 30 · CRESTA · 2025

CLOUDWATCHCROSS-ACCOUNTGRAFANAOBSERVABILITY

One observability plane across thirty accounts.

A logistics company had thirty production accounts and an on-call rotation that toggled between four separate Grafana instances depending on the alert. We unified observability with CloudWatch cross-account sharing and a single Grafana fronted by IAM Identity Center.

INDUSTRY

B2B logistics

DOMAIN

LANDING ZONE

DELIVERED

2025

STACK

CLOUDWATCH·CROSS-ACCOUNT OBSERVABILITY·AMAZON MANAGED GRAFANA·AMP (MANAGED PROMETHEUS)·X-RAY·CLOUDWATCH LOGS

RESULTS

What changed, by the numbers.

ON-CALL TOOL SWITCHES

−100%

PER INCIDENT

MTTR

−42%

LESS TIME FINDING THE METRIC

GRAFANA INSTANCES

4 → 1

MANAGED

LOG INGESTION COST

−27%

CENTRALISED RETENTION

HOW IT WENT

The on-call engineer’s first task during an alert was to figure out which Grafana to open. The second task was to find the right metric — every account had drifted on dashboard layout. The third task was to actually start investigating. Two of the four steps were tooling friction.

CloudWatch cross-account observability turned every production account into a "source" linked to a "monitoring" account. AMP took the Prometheus metrics; Managed Grafana sat on top of both. We rebuilt the on-call dashboard from scratch, with one row per service across all accounts.

On-call drills now start with the right dashboard already on screen. MTTR dropped 42% on the engineering metric, mostly recovered from the "figuring out the tool" step. The four legacy Grafanas were retired in week eight.

READY WHEN YOU ARE

Let's get your AWS bill (and architecture) in order.

The discovery call is free. You walk away with at least one concrete idea — even if we never work together.

Or email directly →