We run the system after launch.
Operate is a retained engagement: we support, maintain, and continue developing your system in your AWS account, under a service agreement renewed annually.
Much of a system’s cost and risk comes after launch: incidents, security updates, infrastructure costs, and the changes the business needs next. Under Operate, the team that built the system takes responsibility for all of it.
What the engagement covers.
On-call cover and incident response
Production issues are detected, triaged, and resolved by the team that built the system, within the response targets in your agreement.
Platform upgrades, security patches, infrastructure maintenance
Security fixes and platform upgrades are tested and applied on a planned schedule, with advance notice of any maintenance.
Quarterly architecture and cost review
A review of reliability, running costs, and architecture each quarter, with recommended changes.
Continued feature delivery against your roadmap
The same team continues to deliver new features against your roadmap after launch.
Air-gapped and GovCloud deployment options
Deployment to AWS GovCloud or isolated environments, where your requirements call for it.
A defined exit, whenever you want to take it in-house
When you choose to bring operations in-house, we transfer them with runbooks, documentation, and training.
How incidents are handled.
Monitoring alerts the on-call engineer when error rates or response times move outside their normal range. Because every request and change is recorded, the cause can usually be identified quickly, and a faulty release can be disabled without a redeployment.
Afterwards, you receive an assessment in business terms: who was affected, for how long, and what will change to prevent a recurrence.
Checkout errors following a release
Illustrative example · Operate engagement
Monitoring alert: checkout error rate above threshold
On-call engineer notified
Cause identified: a change released earlier that evening
Change disabled; checkout restored
Error rate confirmed back to normal
Impact assessment and remediation plan sent to the client
Safeguards from launch.
These are part of the platform, so they are in place from launch whether we operate the system or your team does.
Monitoring on every request
Every request is logged, traced, and measured, so problems can be located and diagnosed quickly.
ObservabilityAudit trail of every change
A tamper-evident record of who changed what, and when, for incident review, audits, and disputes.
Audit LogsPoint-in-time recovery for 35 days
Data can be restored to any point in the previous 35 days if it is affected by an error.
Controlled releases
New features can be released to a subset of users first, and switched off immediately if they cause problems.
Feature FlagsTraffic protection
Rate limits protect sign-in, APIs, and share links from abuse and unexpected spikes.
Rate LimitingPrivacy request handling
Requests to access or delete personal data are verified, processed, and logged in one place.
Data Subject RequestsSet out in your agreement.
Requirements differ from client to client, so the specific levels are agreed in your Operate agreement rather than published here. Every agreement covers the following.
Support hours
Business-hours support as standard, extendable up to round-the-clock cover for critical issues.
Response targets
Response and resolution targets for each severity level, with critical issues prioritized.
Availability
An availability target for the production system, measured and reported against.
Maintenance
Security fixes applied promptly, and upgrades scheduled in advance with notice of any planned maintenance.
Reporting
A quarterly review of reliability, costs, and recommendations, written for business stakeholders.
Incident communication
For any incident that affects your users: what was affected, for how long, and the steps taken to prevent a recurrence.
Operate now, or bring it in-house.
Start when it suits you
Operate can begin at launch or later, and the agreement is renewed annually.
Transfer without a rebuild
When your team is ready, we transfer operations with runbooks, documentation, and training. The system continues to run unchanged.
Continue on the platform
After a transfer, your engineers can keep developing the system on the platform through Self Service.