Fully Managed Ctfreak
as a Service

Operational standard: MSPs typically commit to a monthly patching cadence per client and need strict per-client tenant isolation for credentials, node access, and audit logs. Mixing one client's SSH keys into another client's scope is a contract-breaking event.

DANIAN configuration choice: one Ctfreak project per client. We turn on the Node Set filter and Database Set filter at the project level (the feature added in Ctfreak v1.34, January 2026), so a non-admin user with a role on Client A's project sees only Client A's nodes and credentials. Cross-client visibility is technically impossible, not just discouraged.

Workflow example: a monthly Windows Update playbook task chains to a verification HTTP request task that hits each patched node's status endpoint. On any failure across the fleet, a GitLab notifier opens an issue in the MSP's internal tracker. Each client gets a per-execution PDF report generated by an SQL Report task summarising patched and unpatched nodes.

Quantifiable detail: per-project compartmentalisation is enforced at the data layer — a non-admin user querying the REST API across projects returns 403 on projects they have no role in. The User Information endpoint (POST /userReports) was restricted to admin-only in v1.28 (May 2025), closing a previous information-disclosure path.

Operational standard: teams typically migrate off Rundeck for three reasons — JVM heap consumption (4–8 GB at rest), slow UI on Rundeck Enterprise, and the per-user pricing on PagerDuty Process Automation. The transitional state is delicate: schedules cannot drop, runbooks need to keep firing during cutover.

DANIAN configuration choice: OpenID Connect SSO pre-configured against the team's existing identity provider (Google, Microsoft, Onelogin, Salesforce presets supported in Ctfreak since v1.3). Parameterizable bash script tasks replace Rundeck job definitions; the Smart chaining multiple-execution policy (added v1.8.1) queues new runs rather than dropping them when a previous execution is still active.

Workflow example: migrate one Rundeck job per day to Ctfreak; run both schedulers in parallel for two weeks; cut over after a clean run cycle. Workflow tasks reproduce Rundeck's option-passing pattern via task parameters. GitHub incoming webhooks fire CI runbooks directly from PR comments.

Quantifiable detail: Ctfreak runs in under 2 GB of RAM at idle. JYP's published customer testimonials cite two explicit Rundeck migrations — Karl Leicht (Vigie, France) and Graham R. Brewer (Hometown Computer Services, US) — both naming Rundeck's memory consumption and UI complexity as drivers.

Operational standard: retail ops typically need hourly inventory sync windows, multi-region availability during traffic peaks, and tight cron schedules tied to external API rate limits. Missing a sync window means stale inventory shown to customers.

DANIAN configuration choice: Local Command tasks with per-second cron expression granularity (the optional 6-field cron format added v1.21, October 2024) for tight inventory sync windows. HTTP Request tasks hit Shopify, WooCommerce, or BigCommerce APIs on schedule without writing wrapper scripts. Discord and Slack notifiers fire on failure during business hours; SMS-style escalation routes through Telegram off-hours.

Workflow example: Shopify webhook fires an Ctfreak incoming webhook on order events; a Workflow task chains an HTTP Request to validate stock, an SQL Script to update the inventory warehouse, and a notification to the operations Discord channel. A separate hourly task rebuilds the sitemap and pushes it to a CDN endpoint.

Quantifiable detail: Ctfreak's incoming-webhook task type supports signature verification for five source types — Generic, GitHub, GitLab, Gitea, and Forgejo — with Gitea and Forgejo added in v1.36.1 (February 2026). Per-second cron granularity (e.g. 20 * * * * * for execution every minute at the 20th second) is unique among lightweight schedulers.