+ jHj,Rt^RIHt^RIt^RIt^RIHt^RIHtHtH t ^RI H t H t ]P!R4tRtRt^s`` (Codex 429: the ``AuthError`` from ``hermes_cli.auth_codex._codex_quota_exhausted_error``). When the whole fallback chain is unavailable, re-firing on cadence is guaranteed to fail identically until the window reopens — every fire is a usage probe plus a delivered failure alert. The failing run's alert says the job is held; ``mark_job_run`` then parks ``next_run_at`` at the recovery boundary (or the first legal occurrence after it, when several fall inside the window) and stamps ``quota_hold_until`` so the stale-error re-arm (``cron.jobs._job_is_stale_error_recurring``) does not pull the job back early. Complement to ``cron/unreachable_retry.py``: this one moves ``next_run_at`` out of a known closed provider window. Any run that reaches the model clears the marker. ) annotationsN)datetime)AnyDictOptional)now safe_strftimezcron.schedulerquota_hold_untilquota_hold_cron_exprzretry after (\d+)sc V^8dQhRRRR/#)exc BaseExceptionreturnOptional[float])formats"4/home/ubuntu/.hermes/hermes-agent/cron/quota_hold.py __annotate__r&s=_c^RIHpHp\4pTpVe\ V4V9dVP \ V44\ WA4'dV!V4'du\VRR4pVfC\P\V44pV'd\VP^44MRpVe\V4^8d \V4#R#VP;'g VPpKR#)aSeconds the provider said it will stay closed, or None when *exc* (or anything in its cause chain) is not a rate-limited ``AuthError`` carrying a wait hint. Anchored on the AuthError itself, never on arbitrary text, so an unrelated "retry after" in an agent's output cannot park a job.) AuthErroris_rate_limited_auth_errorN retry_after)hermes_cli.authrrsetidadd isinstancegetattr_RETRY_AFTER_REsearchstrfloatgroup __cause__ __context__)r rrseencurhintms& rhold_seconds_from_failurer+&s FUD#&C /bgT1 C c % %*DS*I*I3 t4D|#**3s84,-uQWWQZ(4"&"2uT{Q5; PD Pmm..s rc$V^8dQhRRRRRR/#)r jobdict[str, Any]rzOptional[datetime]rboolr)rs"rrr;s'MM^M*<MMrc^RIHpHpVP\4'dV!VP\44MRpVRJ;'dT!YA;'g \ 44#)zSTrue while the job is parked inside a provider window (an expired marker is inert).)_instant_after _parse_awareN) cron.jobsr1r2get STATE_KEY _hermes_now)r-rr1r2untils&& r hold_activer8;sJ603 0B0BL+ ,E   L L7K7Kkm!LLrc V^8dQhRRRR/#)r r-r.rNoner)rs"rrrCs%%^%%rc^VP\R4VP\R4R#N)popr5SCHEDULE_EXPR_KEY)r-s&r clear_stater?CsGGItGG t$rc$V^8dQhRRRRRR/#)r r-r.next_runr"rr/r)rs"rrrHs!  . C D rcVPR4;'g/pVPR4R8H;'dHVP\4V8H;'d(VP\4VPR48H#)zTrue for the exact off-lattice cron fire parked by ``plan_hold``. The expression fingerprint keeps a direct ``jobs.json`` schedule edit from inheriting the exception: edited schedules must still re-anchor without firing. schedulekindcronexpr)r4r5r>)r-rArCs&& ris_recovery_firerGHsp wwz"((bH V& ? ? GGI ( * ? ? GG% &(,,v*> >rc V^8dQhRRRR/#)r hold_secondsr#rrr)rs"rrrVsSSeSSrcX^RIHpV!\4\V4\,4#))_seconds_after)r3rLr6r#HOLD_SLACK_SECONDS)rIrLs& r _window_endrNVs( +-|)rEinterval)r3rZr2r[r4r?rNrV isoformatr>r=r5loggerwarningr#) r-rIrWrZr2r[rCrDrQrPblockedrecoverparkeds &&$ r plan_holdrhpsJIwwz"((bH << D ''3777+;x+GC\*J 67Ld"OOol&OGv~GGg+GG2GGG':F 7C v~'!(,@,@,BC]]zG[G[G]%%'!)f!5  !4(  N NN ( c*+U<-@&J rc$V^8dQhRRRRRR/#)r r-r.rIrrr"r)rs"rrrs!  ^ ? s rcV'd0VPR4;'g/PR4R 9dR#\V4p\V4R, pRVR R\VR4 R 2#) zOLine appended to the ONE failure alert delivered on entering the hold, else "".rCrDg @z1 The provider's usage window is closed for about z.1fzh. This job is held through z%Y-%m-%d %H:%M %Zzt and resumes at the first safe opportunity afterwards; no further alerts are sent while the provider is unavailable.>rEra)r4rNr#r)r-rIrQhourss&& r hold_noticermst CGGJ/552::6BJ^^\*J , & (E rMcompile IGNORECASEr r+r8r?rGrNrVrhrmrrrrws # &&9   + ,  ***2BMMB*M% S W(&OT&R r