Grafana monitoring and alerting
Grafana is where the numbers become a decision. A dashboard that answers "is it the database or the API" in ten seconds is worth more than any amount of collected data nobody looks at.
Where we use Grafana
Grafana is part of the stack on these 2 services. Each page covers how we work, what you get and what it costs to start.
Grafana in practice
Grafana is where collected data becomes a decision. The measure of a dashboard is not how much it shows but how quickly it answers the question you have at 3am: is it the database, the API or the upstream provider? A dashboard that takes ten seconds to answer that is worth more than twenty that look impressive.
We build a small number of dashboards deliberately: latency, error rate, saturation and throughput per service, plus the business metrics that tell you whether users are actually affected. Definitions live in version control, so a dashboard does not disappear with whoever created it.
What we build with Grafana
Latency, error rate, saturation and throughput on one screen, so a page can be triaged before anyone opens a log.
Signups, orders and conversion alongside technical metrics, because they are usually related.
Alerts on symptoms users would notice, tuned so that being paged still means something.
Is Grafana right for you?
Ask usA good fit when
- Visualising Prometheus, cloud and log data in one place
- On-call teams needing fast triage
- Business and technical metrics viewed side by side
- Organisations wanting control and predictable monitoring cost
Probably not when
- Teams with nothing collecting metrics yet — start there first
- Organisations that would rather buy a fully integrated product
- Cases where nobody has time to maintain dashboards
What we run alongside Grafana
The rest of the setup, and why each piece is there. We keep this list short on purpose — every dependency is something someone has to maintain.
- Prometheus
- The usual metrics source, and the pairing the ecosystem assumes.
- Loki
- Logs queryable alongside metrics, so correlation does not mean switching tools.
- Alertmanager
- Routing, grouping and silencing, so alerts stay meaningful.
- Dashboards as code
- Definitions in Git, provisioned automatically rather than clicked together.
- SLO panels
- Error budgets, so alerting is tied to what users experience.
Why Grafana
Let’s talkReads from anything
Prometheus, cloud metrics, logs and SQL databases in one place, so correlation does not mean switching tools.
Dashboards as code
Definitions stored in version control instead of being clicked together and lost with an account.
Alerting built in
The same thresholds you watch on a dashboard become the alerts that wake someone up.
What we get called in to fix
Get a second opinionDashboards nobody uses
Dozens of panels showing everything and answering nothing.
Alert fatigue
Alerts on causes rather than symptoms, until the team routinely ignores them.
Dashboards lost with an account
Definitions clicked together and never exported, gone when someone leaves.
No business metrics
Perfect technical dashboards that cannot tell you whether customers are affected.
Grafana or the alternative
The comparisons we are actually asked to make, answered the way we would answer them on a call.
The open pair for control and predictable cost. Datadog when you would rather buy the integration work and correlation.
Cloud to avoid operating storage. Self-hosted when data residency or cost at volume decides it.
Loki is cheaper and simpler for log aggregation. Elasticsearch when you need full-text search across logs.
Grafana works well with:
AWS
Azure
Docker
Google Cloud
Kubernetes
Prometheus
Python
Redis
Terraform
Angular
AnsibleGot an idea? Let’s make it real.
Tell us the short version
This could be the first step towards a new and successful collaboration. A one-line idea and a finished spec are both fine — tell us the problem, the deadline you’re working to and what’s in your way.
Keep looking
Frequently asked questions
Yes, and we walk through them at handover. A dashboard nobody understands is not monitoring.
Grafana and Prometheus when you want control and predictable cost; Datadog when you would rather buy the integration work.
Yes — Slack, email, PagerDuty and Opsgenie are all standard integrations.
Few, and used. One per service for triage, one overview, and business metrics. Everything else tends to be noise.
Yes — alerts on user-visible symptoms, routed to Slack, PagerDuty or Opsgenie, and tuned so being paged still means something.