Who it is for
A product with paying users that learns about outages from support tickets. Either there are no alerts, or there are so many noisy ones that everyone has muted them.
The problem
- There's no agreed definition of 'working', so alerts fire on CPU instead of on user pain.
- Noisy alerts get muted, and real incidents get missed.
- No runbooks: every incident starts from zero, at night, with the person who knows the least.
What I do
- Dashboards for traffic, errors, latency and saturation for each service.
- SLOs agreed with you (for example 99.5% availability), with recording rules and multi-window, multi-burn-rate alerts that are unit-tested.
- Uptime checks from outside, and routing by severity to Slack or a pager.
- A runbook for every alert, rehearsed in a game day.
How it works
- 1Metrics and logs from each service
- 2SLOs agreed with you
- 3Burn-rate alerts (tested)
- 4Routing to Slack or a pager
- 5A runbook per alert
- 6Game day to prove it
Day by day
- Day 1-2Inventory of services and signals; dashboards for the golden signals.
- Day 3SLOs agreed with you; recording rules and burn-rate alerts with unit tests.
- Day 4-5Uptime checks, routing, and runbooks for every alert.
- Day 6-7A game day: we inject a failure and follow the runbook; postmortem and fixes.
Evidence from the demo
The evidence below comes from my own demo project with synthetic data. It shows exactly what you receive; it is not a story about a client.
- KubeOps SRE Lab runs Prometheus, Grafana, Loki and Alertmanager on a laptop with one command, all deployed by Argo CD from Git.
- SLOs of 99.5% availability and 99% of requests under 300 ms, with burn-rate alerts unit-tested by promtool.
- Four scripted game days (error spike, latency regression, pod or node failure, blocked insecure deploy), each with a runbook and a postmortem.
Demo project: KubeOps SRE Lab
What you keep
- Dashboards and alert rules as code
- SLO document with error budgets
- Runbooks and routing
- Game-day report
Not included
- Being on call for you
- Migrating to a different monitoring vendor
Questions
Do you need access to my production systems?
No passwords, ever. Audits use a read-only role you create and can delete at any time; changes arrive as a pull request you review and merge yourself.
How long does a project take?
2 to 10 working days depending on the package. You get the timeline and the fixed price before we start.
What if you find nothing worth fixing?
Then you hear that early, with the evidence. For the AWS Bill Doctor on accounts above $1,000 a month: if I don't find at least my fee in yearly savings, you don't pay.
What happens to my data?
Reports stay between us. When the project ends I delete your data and confirm it in writing.
How do I pay?
Upwork, international bank transfer, or UPI / bank transfer in India. Direct clients pay 50% upfront and 50% on delivery.