00:00:00 UTC
sevenroute.dev
Field notes ◆ Linux · AWS · Observability · Automation Vol. 04 / Issue 12 Latest: Tuning Logstash pipelines for cost Status: shipping Available for new infra work — Q3 2026

Infrastructure,
observability &
the quiet art of
keeping things up.

06
Years in production
99.94%
Uptime kept, last 12 mo.
~1.2TB
Logs ingested daily

Featured work

See all projects →
Case 001 — 2025 EC2 / ELK / Grafana
┌──────────────┐ │ ec2-prod-1 │ ▓▓▓▓▓░░ 62% │ ec2-prod-2 │ ▓▓▓░░░░ 41% │ ec2-prod-3 │ ▓▓▓▓▓▓▓ 88% └──────────────┘ logstash → es → grafana
Watching Linux,
from a long way off.
— Featured · Case study

Monitoring Linux on AWS EC2 with the ELK stack & Grafana.

A self-hosted observability pipeline for a fleet of EC2 instances. Logstash collects from Filebeat, Elasticsearch indexes by service, and Grafana surfaces SLOs alongside system metrics.

Built to be cheap, restartable, and quietly opinionated. Total monthly cost stayed under $40 by tuning ILM policies and using spot-priced data nodes.

Terraform Elasticsearch Logstash Filebeat Grafana EC2 / Spot
Read the case study →

Recent writing

All entries →

Now,
in April 2026

Building

A small Terraform module for self-hosted Grafana that fits in one EC2 + S3.

Reading

Brendan Gregg on system performance, again. Different chapter, same coffee.

Listening for

SRE / platform roles that take observability seriously. Get in touch.