--- title: "Backlog — StatHall" description: "StatHall product roadmap: the V2 'universal cockpit' delivered in independent functional lots — mobile apps, Unix servers, and beyond." lang: en canonical: "https://www.stathall.com/en/backlog/" --- Roadmap # Product roadmap The V2 "universal cockpit" ships in **independent functional lots** — each one shippable alone, with no imposed date milestone. The direction is clear, and so is the order. For the release-by-release detail, see the[public changelog](https://www.stathall.com/en/changelog/). ## In progress The lot being built right now. - ### Publishing mobile app 3.2 mobile 3.2.1 The app is rebuilt and tested; what remains is publishing it to the App Store and Google Play. The latest fix closes the leak that mattered most: switching accounts now reloads every screen, even when the app was closed without signing out. - ### Wiring the last shipped functions that have no screen Fourteen API operations exist without a dedicated page in the cockpit — each is measured, with its reason, and they disappear as the screens land. ## Planned The direction of the V2 "universal cockpit", delivered in independent functional lots. No firm date, but a clear order. - ### Lot 6 — Jira, GitLab and GitHub connectors The three connectors currently in preview become functional: issues, sprints, merge requests, pipelines — brought next to crashes and reviews. Available on apps as well as servers. - ### Lot 7 — Zero-downtime deployment Industrialised releases: instrumented rolling restart to cut the unavailability window to under one second during releases. - ### Lot 9 — Automatic agent updates (optional) A signed distribution channel so installed agents update cleanly, at the pace chosen by the administrator. - ### Beyond — databases, websites, containers The polymorphic base laid in Lot 1 is built to host new target types without a rewrite. Databases, e-commerce sites and Docker containers are the logical continuation of the universal cockpit. ## Released Already in production. - ### Export your data to Excel or CSV v3.11.0 Monthly trend and Sentry quota beyond Sentry's 90 days, crash-free by release, stores, 400 days of server metrics: an Exports screen per target and for the fleet, ready for a pivot table. - ### Every alert says what and how much v3.8.0 A gauge, a figure or a name for each of the 38 alert types, with the icon of the resource concerned; diagnostic leads specific to each resource. Cockpit and phone. - ### “CI runner” profile and job containers ignored v3.8.0 A continuous integration machine no longer alerts on every job: a profile to tick, suggested when GitLab Runner runs; ephemeral job containers are ignored everywhere. - ### Filter the inbox by type, group and target v3.7.2 A server's anomaly badge leads to its own anomalies, no longer to the whole fleet's. The inbox can be filtered by type, by group (sub-groups included) and by target, on the cockpit as on the phone. - ### What fixed itself is still there to see v3.1.3 An alert e-mailed at 3:30 and fixed by the server at 6:00 vanished from the inbox: it now stays there for twenty-four hours, marked “Settled” and saying what fixed it. Two alerts that taught nothing also went quiet: a full swap on a machine that never swaps, and the noise of a build server. - ### One alert per incident, and nothing left from the previous account v3.1.2 A container restarting in a loop produced fifty-four alerts for a single outage, and bots turned away by a web server were counted as application errors. Both are quiet now. And switching accounts no longer leaves the previous one's data on screen, even when the session ended without a sign-out. - ### Alerts that stop crying wolf v2.9.1 A CPU alert fired on a single busy core of a sixteen-core machine, and a one-minute spike opened a quarter-hour alert. Alerts now look at the machine, require the breach to hold, and say whether it is a spike or sustained. Every anomaly offers where to look and a ready-to-copy prompt for the AI assistant of your choice. - ### One menu, and nothing hiding any more v2.9.0 Five competing menus merge into one. Creating a target is no longer reserved to administrators. One screen gathers the AI keys of all your targets. And a site map lists every screen, including those your permissions hide — with the reason. - ### Predictive monitoring, alert channels and vulnerability watch v2.8.0 → v2.8.7 · agent 1.3 Disk saturation announced ahead of time, alerts over e-mail, push, signed webhook and Telegram, software inventory with security advisories, and an alert brief you can act on without opening the cockpit. - ### Resilient logs and log-derived detection v2.7.0 → v2.7.8 · agent 1.2 Bounded log reads with stable pagination, a version-tolerant ingestion protocol, and seven new anomaly types derived from logs: error bursts, disk errors, SSH brute force, crashed processes. - ### Mobile app 3.0 — new generation mobile 3.0.0 The iOS and Android app rebuilt on a fresh foundation: tablet-aware layout, more reliable notifications, offline mode that replays your actions when the network returns. - ### Mobile 2.5 — offline, background refresh, rich notifications mobile 2.5.0 Persistent cache and offline reading, actions (acknowledge, resolve, snooze) queued and replayed when the network is back, background refresh with badge, notifications with title, body and an "Acknowledge" action, overview, app ↔ server links, server target creation with an install QR code, alert thresholds, probes and containers on the server page. - ### Lot 3 — Overview, app ↔ server links and real-time cockpit v2.6.0 Overview of every target sorted by priority with the "hosted on" topology, real-time updates (no more periodic polling), paginated and smooth inbox, cockpit installable as an app, four times lighter initial load. StatHall monitors itself (scheduler, workers, connectors, data freshness). - ### Lot 2 — End-to-end server monitoring: agent 1.1 + probes v2.5.0 · agent 1.1.0 Unix agent 1.1 (twenty system metrics, Docker containers, systemd timers, reboot required, out-of-memory kills), HTTP / TCP / TLS probes run from the cockpit with expiring-certificate alerts, silent agent flagged as an anomaly, per-server thresholds, bounded retention. Reliable app detection: Android ANR rate ingested and notified, latest available data point used, escalations notified. - ### Lot 4 — Fair-use, donations and open-sourcing v2.4.0 No more pricing or subscription: free hosted instance (fair-use), code published as open source (AGPL-3.0) for self-hosting, voluntary support through donations. - ### Lot 1 — The "universal cockpit" base v2.0.0-rc.1 Internal rework of the data model: StatHall no longer supervises just "apps" but generic "supervised targets", ready to host servers (Lot 2) then databases, websites and containers. Transparent migration — every V1 app, connector and history is kept identical. Historical Buy Me A Coffee contributors are recognised as "Founding supporters". Health-score reliability fixes in rc.2 → rc.5 are tracked in the public changelog. - ### Before V2 — StatHall V1 v1.0.0 The whole inherited base: mobile app supervision (Sentry, Apple App Store Connect, Google Play, Firebase), AI review summaries, anomaly detection, crash ↔ review correlation, roadmap, multi-tenant SaaS mode and self-hosted mode, iOS/Android mobile app, marketing site. The release-by-release detail remains available in the public changelog. Got a request or a blocker we haven't covered?[Write to us](https://www.stathall.com/en/contact/)— the roadmap is alive, not carved in stone.