Site reliability engineering for important applications

Measure the experience users depend on

A server can be running while an application fails to complete an important task. A slow checkout, delayed background job or unreliable data import may matter more to the organisation than a headline uptime figure.

Our site reliability engineering work starts with those user journeys and the consequences of failure. We help define the indicators worth measuring, the reliability objectives that fit the service and who will review the evidence. The objective should inform an engineering decision, rather than exist only on a dashboard.

Connect monitoring to a response

We review the application, infrastructure and dependencies to understand where failures originate and what the current monitoring reveals. Logs, errors, request times, queues and scheduled jobs can each provide part of the picture.

Alerts need clear thresholds, context and an owner. We work with your team to distinguish failures requiring action from information better reviewed during planned maintenance. Support coverage and escalation are agreed separately from the monitoring configuration; an alert does not create a response arrangement by itself.

Prioritise the changes that address recurring risk

Reliability work may involve database queries, capacity, background processing, dependency failures or the way changes reach production. We investigate before deciding which part of the system needs to change.

The resulting plan can include:

• strengthening tests around a recurring fault or critical journey;
• making retries and failure handling appropriate to the operation;
• improving backup, recovery or rollback procedures;
• removing repetitive operational work through controlled automation; and
• documenting runbooks and the ownership of unresolved risks.

We balance these tasks with the product roadmap and review whether the changes improve the agreed measures. Reliability work is ongoing prioritisation, not a promise that incidents will disappear.

Relevant experience with live applications

For Easol, we investigated application performance and hosting, including database queries and how pages were rendered. The work helped halve page-load times for some customers. For Serious Readers, we improved an underprovisioned hosting setup and reworked an unreliable ERP integration.

Each project started with specific constraints. The same change would not necessarily be appropriate for another application, even when the symptoms look similar.

SRE, DevOps or ongoing support?

This service concentrates on how reliably an application performs its job in production. Our DevOps consulting focuses on environments, testing and the delivery process. Support and maintenance establishes continuing application ownership and coverage.

An initial discussion covers the service, operational evidence and recent incidents. We then agree the assessment, responsibilities and improvements that fit your team’s capacity to operate the system.