The Disciplines Plate VI Monitoring & Observability
Monitoring &
Observability.
So you hear first.
Dashboards, alerts and telemetry set up so that when something goes wrong, the first person to hear about it is you, with time to fix it before anyone else notices. Prometheus, Grafana, Zabbix, Checkmk.
Return to the workshopSome businesses can't send monitoring data to the cloud for compliance or security reasons, and need it kept on site, sometimes fully air-gapped. Others are happy with a hosted service. Most end up somewhere in between.
I design the monitoring around your setup, whether that is a handful of servers or hundreds of endpoints, and balance what you need to see against security and cost. Where I can, I use open-source tools (Prometheus, Grafana, Uptime Kuma, Checkmk), which do the job of the big commercial platforms without the licence fees. The money goes on setting it up, and none of it on software.
§ What I monitor
-
I.
Infrastructure health
CPU, memory, disk and network. Alerts fire before a server runs out of room, and the history shows when an upgrade is coming, long before it turns into an emergency.
-
II.
Uptime & availability
Your site checked from several locations, with warnings before SSL certificates expire, DNS checks, and response times tracked for SLA reporting.
-
III.
Application & database
Query performance, connection pools, replication lag and slow queries, plus metrics specific to your application. Where the standard tools don't cover something, I add custom instrumentation.
-
IV.
Security events
Failed logins, firewall blocks and unusual network traffic, fed into a SIEM if you have one, and correlated across systems so a pattern shows up even when each event looks harmless on its own.
-
V.
Log aggregation
Logs from every system in one place, searchable across months in seconds, and structured so problems are quicker to trace. How long they are kept is set by what compliance needs and what storage costs.
-
VI.
Alerting
Escalation policies and on-call schedules, delivered by Slack, Teams, email or SMS. Thresholds are tuned to cut false alarms, because people stop reading alerts that cry wolf.
§ The process
-
Step 1
Discovery
A look at what you have and what matters most: which systems are critical, how much downtime you can live with, what compliance applies, and whether it should run in the cloud, on site or air-gapped.
-
Step 2
Design
A proposal that fits your requirements and budget, saying which tools to use, where they run and which metrics are worth tracking.
-
Step 3
Deploy
Agents installed, dashboards and alerts set up, and everything connected to the tools you already use, with notes on what was deployed and why.
-
Step 4
Tune
The first month is for adjusting: thresholds moved, noisy alerts cut, metrics I missed added and useless ones dropped.
-
Step 5
Support
Ongoing support if you want it, adding new systems as they arrive and helping when something goes wrong. If you would rather run it yourself, you get a clean handover.
Questions, answered
We already get alerts and ignore them. How is this different?
If people ignore the alerts, the alerts are badly designed. Most of the work is deciding what is worth waking someone for and deleting the rest, which usually leaves you with fewer alerts than you have now.
Do we need to buy a monitoring platform?
Usually not. For a business of this size, the open-source tools named above do everything the licensed products do, and the money is better spent on having them set up well.
What will it cost?
The prices on this page are starting points, so you can tell in a minute whether you are in the right range. Each job is then quoted after a conversation, because the price depends on the shape of the job and nobody knows that yet. The first half hour is free, and it usually settles whether this is a small piece of work or a large one before any money is discussed.
Who owns the code when it is finished?
You do, outright, including the source. It is handed over with the runbooks needed to keep it running, and there is no licence to renew and no seat to pay for. Where this page quotes a monthly figure, that pays for hosting or ongoing support, which is optional and can be stopped without taking anything away from you. Commissions, not subscriptions, is meant literally.
What happens if you are unavailable?
Everything is built so somebody else could pick it up: common technologies, readable code, written runbooks and nothing proprietary. It is a fair question to ask of a one-person business, and the answer is in how the work is built.
Do you work outside Suffolk and Cambridgeshire?
Yes. Local work gets the option of somebody in the room, which is worth more than it sounds for the first conversation and the handover. Everything after that is done remotely for most clients anyway.