v2.9.1
A patch release for the Cloudflare solver stack. If you're on 2.9.0 and your Networks Top 10 collections came back empty with no explanation, this is the release that tells you why and gives you the tools to fix it. You can now run more than one solver (FlareSolverr, Byparr, or both) with automatic failover, and a new health check flags the setup gap instead of letting collections fail silently.
Docker tags
| Tag | What it tracks |
|---|---|
bitr8/agregarr:latest
| Stable releases. This is what you want. |
bitr8/agregarr:2.9.1
| Pinned to this exact release. |
bitr8/agregarr:develop
| Bleeding edge. Builds on every push, breaks sometimes. |
Upgrading from 2.9.0 is a pull. An existing FlareSolverr URL migrates into the new solver list automatically; nothing to reconfigure.
What's new
Solver priority list with failover
Settings > Sources now takes a list of Cloudflare solvers instead of a single URL. Add FlareSolverr and Byparr instances side by side, name them whatever you like, and drag to set priority. Fetches try each entry in order and the first one that answers wins.
Backoff is per instance, per domain. A solver that keeps failing gets skipped for a while without dragging the others down, so a flaky instance costs you a retry, not your collections.
The connection health check probes every instance in parallel and names the dead ones: all down is an error, some down is a warning.
Changed
Plain HTTP for FlixPatrol is gone
Cloudflare's challenge now blocks plain HTTP against FlixPatrol outright, so the 2.9.0 toggle only produced broken configs. Removed. FlixPatrol fetches always route through the solver stack now: your configured solvers, or the built-in Chromium when none is set.
Alongside that, a new health check. If you have networks collections and no solver configured, the Health panel says so and points at FlareSolverr or Byparr setup. The silent version of this failure was the worst part of running 2.9.0 without a solver: the collection just came back empty and nothing told you why.
Fixed
- The solver health probe works with Byparr. It used
sessions.list, which is FlareSolverr-specific; a plain GET to the base URL checks reachability without assuming the API surface. (fork #61) - Built-in browser failures no longer block a freshly configured solver. Backoff was keyed by domain alone, so a few Chromium failures before you set a solver URL kept the new solver from being tried once you had. Superseded by the per-instance backoff above, but worth knowing if you saw it on 2.9.0.