What OneUptime/oneuptime shipped

This week: error budget policies and a slow burn rule for SLOs, metrics on root cause pages for incidents, alerts and episodes, evaluation details showing why a monitor fired, and telemetry usage per project and service on every install.

This week in OneUptime: SLOs can carry an error budget policy with tiers and alerts when an SLO enters one, and every new SLO gets a slow leak burn rate rule. Root cause pages for incidents, alerts and episodes now show the metrics behind them, evaluation details show what a check found with an option to keep no response content, and telemetry usage is counted per project and per service on every install, billing on or off. Approvers can decide on waiting public notes from Slack, Teams and Mattermost, and workflow saves now ask whose version to keep instead of overwriting a teammate's changes.

This week: Kubernetes and Proxmox views with new alerts, App Platforms with deploy history, exception spikes judged against the hour of the week, metric anomaly fixes, and sessions that keep your theme and language.

This week's releases bring several changes for people running OneUptime day to day. Kubernetes gets a page for every workload, pod and node with shared health rules and three new alerts, and Proxmox gets an attention-first view with per-node pages, backup results and five alerts. App Platforms now lists every app with a page, deploy revisions, rollbacks and per-app requests, errors and latency. Exception spikes are alerted against what is normal for the hour of the week, and metric anomaly checks now read the right weekday. Status page messages to subscribers are written in their own language, and a session ending no longer clears your theme, language or cookie choice.

14.0.90 is out: connect GitLab, Forgejo/Gitea and any Git repository beside GitHub, Meraki traffic beside NetFlow, and enterprise backups for PostgreSQL and ClickHouse. Plus Entity Risk in Security Events and an easier self-hosted setup.

Release 14.0.90 is out. You can now connect GitLab, Forgejo/Gitea and any Git repository beside GitHub, bring Cisco Meraki dashboard traffic in beside NetFlow, and on Enterprise back up PostgreSQL and ClickHouse to your own storage on a schedule with safe restore. Security Events gained Entity Risk for risk-based alerting, project-owned case priorities and statuses, and security cases in the On-Call mobile app. Self-hosted setup now checks the server, makes every secret and tells you where to open OneUptime, and a deploy under an open tab shows a quiet notice instead of an error.

This week: Zabbix workflow templates, infrastructure agent installs via apt, yum, Homebrew and winget, replicas spread across nodes in Helm, HTTPS-only sign-in cookies, per-IPv6 /64 rate limits, and no more default cluster keys.

This week's changes to OneUptime: the Create a workflow dialog now offers Zabbix templates that open incidents or alerts for Zabbix problems and send acknowledgments and notes back. The infrastructure agent installs through apt, yum, Homebrew and winget, and its Windows MSI upgrades in place. Helm charts spread every stateless service's replicas across nodes. Sign-in cookies stay on HTTPS and HTTPS installations tell browsers to stay there. Rate limits keyed on client addresses count each IPv6 /64 as well as each address, and wrong two-factor codes are counted for the account once the password is right. The cluster key is no longer accepted if it is a key the repository ships.

This week: one complete walk for markup across test files and product code, Jira and Dynamics text read as written, and acknowledging alerts on declare matches acknowledging them anywhere else. Plus docs rewritten and translated in every language.

This week OneUptime took a step back on text handling: test suites and product code that take markup out of text now use one shared complete walk instead of a copied pattern per file, so rendered pages read as the words a person sees. Text from Jira and Dynamics 365 reads as written in incident and alert titles, descriptions and notes, Slack webhook payloads stay inside their code block, and Teams pinned messages are read in one pass. Acknowledging linked alerts while declaring an incident now takes the same as acknowledging them one by one, episodes move through states by the incident and alert rule, and scheduled maintenance starts, rolls up and shows completed by where its state sits. The daily retention job hard deletes network sites and site types deleted more than 30 days ago, Jira i

14.0.14: Forms now start on what matters, search covers settings, and custom branding works on public forms. We're making OneUptime simpler to use, one form at a time.

14.0.14 ships this week with a focus on form simplicity across the product. Stepped forms now show their primary action only on the last step. Label rules, owner rules, and metric pipelines start directly on configuration instead of a Basic Info step. Search (Cmd+K) now indexes settings pages like API Keys and Delete Project. Custom branding on public forms is here. We're systematically reducing decision paralysis to make OneUptime less tiring to use.

14.0.7 ships server-side topology, email verification resend, status-page scoped incidents, and security fixes for DNS-rebinding and egress leaks.

14.0.7 moves topology computation to the server so Device Topology search works across the whole hierarchy, redesigns email verification with resend, scopes incidents to status pages with targeted notifications, and closes security gaps in DNS-rebinding and egress filtering for SaaS tenants.

13.0.7 shipped: Enterprise Edition under its own license, IPv6 monitors fixed, on-call timeline view across schedules, Security Events volume chart, and multi-tenant permission enforcement improvements.

This week's releases include major structural changes: Enterprise Edition moved to its own ee/ directory under separate licensing, keeping community images focused on open source. Operationally, on-call schedules now support unified week/month timeline views, Security Events gained severity volume charting, and we fixed IPv6 monitor destinations. On security, we hardened multi-tenant authorization checks and permission block enforcement to prevent sensitive data exposure.