2026-06-11 — Reliability Upgrades: Sturdier Background Jobs and Failover
Not every update is a feature. Some of it is the plumbing — the work that doesn't change what you see but changes whether the thing keeps working at 3am. This is a roundup of that.
Highlights
- Background jobs moved onto a dedicated, proven scheduler
- Fixed crashes that were taking the background worker down
- Telemetry lifecycle and provider failover hardened
- Replay and recovery paths made more robust
- Runtime kept current with the latest patch releases
Sturdier background work
The scheduled and behind-the-scenes jobs Momor runs were moved onto a dedicated, battle-tested scheduler, and we fixed a set of crashes that had been knocking the background worker over. The work that's supposed to happen quietly now actually happens quietly.
More resilient under failure
We tightened how telemetry starts up and shuts down, and hardened the failover that moves a request to the next provider when one stumbles. We also shored up the replay and recovery paths, so a search that gets interrupted is more likely to pick itself back up cleanly.
Kept current
The underlying runtime was moved up through its latest patch releases, so we're sitting on the most recent fixes rather than trailing them.
Why This Matters
Reliability is the feature you only notice when it's missing. The payoff isn't a new button — it's fewer stalls, fewer silent failures, and a system that recovers on its own instead of leaving you staring at a spinner.
What this means for you: nothing new to learn, and ideally nothing to notice — just a search that's quietly more dependable than it was a couple of months ago.