Following our beta-tested 1.1.0 release, we’ve published v1.1.1 of ws-availability — our new stable version, with the improvements listed below.
Highlights
- Built-in scheduler — the availability-view refresh now runs inside the cacher container via APScheduler.
avail-rebuild— rebuild the availability view for any historical date range and optional network/station/location/channel, run inside the cacher. Lets you repair older gaps after a WFCatalog backfill (the daily scheduler only covers a recent window).- Single-source configuration —
config.pyis now the one place to configure the service - Optional parallel MongoDB fan-out — large time-window queries can be split across workers for faster responses. Off by default; opt in via config.
- Per-node Sentry tagging —
SENTRY_ENVIRONMENTlets you distinguish nodes/environments in error reporting.
Major fixes since v1.0.0
Response correctness
- Continuous traces spanning multiple days are no longer split across lines/days (#23, #48).
- Spurious and duplicate timespans are no longer returned (#24, #27).
orderbynow works as expected (#25).merge=overlapnow works correctly (#26).- Fixed bad output for simple requests (#40).
Restriction handling
- Restricted stations are no longer reported as open,
includerestrictedis now honored, and restriction status is properly persisted (#39, #33, #55). - Empty location codes are shown as
--in text output (#38).
Performance & stability
- Large requests are handled safely without exhausting memory (#60).
- Major performance improvements on slow availability requests (#51, #57).
- Eliminated MongoDB connection churn under load (#47).
Configuration
- The authentication database is now configurable instead of hard-coded (#18).
Get it
Installation and upgrade instructions are in the README. If you are upgrading from v1.0.x, review your config.py against the updated sample — it is now the single configuration source.
Feedback and issues are welcome on the GitHub.
