Safe Haskell	None
Language	GHC2021

Ecluse.Cve.Sync

Description

The advisory-sync plan: one CveSyncHandle per mount ecosystem (planCveSync), the projections the composition root reads off it, and one supervised sync task per handle.

Synopsis

data CveSyncHandle = CveSyncHandle {
- csReady :: TVar Bool
- csEnv :: SyncEnv
- csMaxAge :: MaxAdvisoryAge
- csClock :: IO UTCTime
- csAgeAlarmed :: TVar Bool
- csOutage :: TVar OutageState
- csDatabase :: DatabaseRequirement
}
data AdvisoryNeed = AdvisoryNeed {
- anEcosystem :: Ecosystem
- anMaxAge :: MaxAdvisoryAge
- anEpss :: EpssRequirement
- anDatabase :: DatabaseRequirement
}
planCveSync :: LogEnv -> Maybe AwsEndpoint -> AppConfig -> [AdvisoryNeed] -> IO (Map Ecosystem CveSyncHandle)
sweepStaleTemps :: LogEnv -> FilePath -> IO ()
sweepStep :: LogEnv -> FilePath -> IO () -> IO ()
cveRuleDepsFor :: Map Ecosystem CveSyncHandle -> BreakerReporter -> (Ecosystem -> OutageReport -> IO ()) -> Ecosystem -> RuleDeps
advisoryFreshnessFor :: Map Ecosystem CveSyncHandle -> Ecosystem -> IO AdvisoryFreshness
reportPushAge :: LogEnv -> Ecosystem -> CveSyncHandle -> IO ()
katipOutageReporter :: LogEnv -> Ecosystem -> OutageReport -> IO ()
outageReportPeriod :: NominalDiffTime
cveSyncReadiness :: Map Ecosystem CveSyncHandle -> IO Readiness
cveSyncScheduleFor :: AppConfig -> SyncSchedule
cveSyncTasks :: LogEnv -> Metrics -> Telemetry -> SyncSchedule -> Map Ecosystem CveSyncHandle -> [IO ()]
registerAdvisoryAges :: Metrics -> Map Ecosystem CveSyncHandle -> IO ()

Documentation

data CveSyncHandle Source #

One configured ecosystem's advisory-sync wiring.

Constructors

CveSyncHandle

Fields

csReady :: TVar Bool
The one-way first-sync readiness flag.
csEnv :: SyncEnv
The sync task's environment. Its syncSlot is the slot this ecosystem's rules borrow through.
csMaxAge :: MaxAdvisoryAge
This mount's effective maximum push age, derived once at boot from its own rules.
csClock :: IO UTCTime
The wall clock the push age is read on, injected so a suite can fix it.
csAgeAlarmed :: TVar Bool
Whether the half-maximum crossing has already been reported for the current push.
csOutage :: TVar OutageState
The rules-side outage state every mount of this ecosystem reports through.
csDatabase :: DatabaseRequirement
Whether this mount's own rules deny on the database, which is what its readiness turns on.

data AdvisoryNeed Source #

What one vetted mount's rules ask of the advisory stack, read off its own policy at boot.

Constructors

AdvisoryNeed
Fields anEcosystem :: Ecosystem anMaxAge :: MaxAdvisoryAge anEpss :: EpssRequirement anDatabase :: DatabaseRequirement

planCveSync :: LogEnv -> Maybe AwsEndpoint -> AppConfig -> [AdvisoryNeed] -> IO (Map Ecosystem CveSyncHandle) Source #

Build the advisory-sync plan, one CveSyncHandle per vetted mount ecosystem, or nothing with no store. A mount the build does not ship awaits an artifact that never comes, so it stays unready.

sweepStaleTemps :: LogEnv -> FilePath -> IO () Source #

Sweep the in-progress downloads an interrupted run left behind, which an emptyDir keeps across a container restart. The sweep is best effort, per sweepStep.

sweepStep :: LogEnv -> FilePath -> IO () -> IO () Source #

Run one best-effort step of the stale-temp sweep. It logs and swallows an IOError, so a read-only or mispermissioned data dir does not stop the boot, and any other exception propagates.

cveRuleDepsFor :: Map Ecosystem CveSyncHandle -> BreakerReporter -> (Ecosystem -> OutageReport -> IO ()) -> Ecosystem -> RuleDeps Source #

The rules' boot-bound capabilities for one mount ecosystem. A mount's rules read only their own ecosystem's advisory database. One the plan carries no handle for has none configured, and reports nowhere.

advisoryFreshnessFor :: Map Ecosystem CveSyncHandle -> Ecosystem -> IO AdvisoryFreshness Source #

How old one mount's serving artifact's push is. An ecosystem the plan carries no handle for has no advisory stack at all, so nothing ages and the absent-database path decides instead.

reportPushAge :: LogEnv -> Ecosystem -> CveSyncHandle -> IO () Source #

Report one ecosystem's push age when it passes half its maximum, once per crossing. The latch re-arms when a fresh push brings the age back under, so a long outage does not repeat every poll.

katipOutageReporter :: LogEnv -> Ecosystem -> OutageReport -> IO () Source #

Log one ecosystem's advisory-source outage reports: the start and each reminder at ERROR, the level an operator pages on, and the recovery at INFO. A fault's detail rides along and reaches no client.

outageReportPeriod :: NominalDiffTime Source #

How often a continuing outage reminds the operator: the unloaded-database report's own gap, so an outage costs the log one line per interval on either path.

cveSyncReadiness :: Map Ecosystem CveSyncHandle -> IO Readiness Source #

The readiness verdict over the sync plan. Only a mount whose rules deny on the database waits for its first sync, and one ecosystem's missing artifact leaves the others routable.

cveSyncScheduleFor :: AppConfig -> SyncSchedule Source #

The sync tasks' timing: the shipped boot burst over the configured poll interval. The microsecond conversion cannot wrap: the config decoder bounds the interval to [1, maxBound div 1_000_000] seconds.

cveSyncTasks :: LogEnv -> Metrics -> Telemetry -> SyncSchedule -> Map Ecosystem CveSyncHandle -> [IO ()] Source #

One supervised sync task per configured ecosystem, each flipping its own one-way readiness flag once its first sync lands. Every role that evaluates rules runs these.

registerAdvisoryAges :: Metrics -> Map Ecosystem CveSyncHandle -> IO () Source #

Register once per role. Callbacks read the slots, so observations survive sync-task restarts.