Commit Graph

1199 Commits

Author SHA1 Message Date
Joseph Doherty e4136cb920 fix(transport): transport template NativeAlarmSources — imports no longer amputate native alarm mirrors nor dangle instance overrides
New TemplateNativeAlarmSourceDto + init-only TemplateDto.NativeAlarmSources (empty
default, additive — IsInherited placeholders carried for the collision detector).
Exported in ToBundleContent, consumed in FromBundleContent + BuildTemplate, and a
new SyncTemplateNativeAlarmSourcesAsync runs the Overwrite add/update/delete child
sync with per-change audit rows. ArtifactDiff.CompareTemplate now DiffChildren over
native sources so Preview reports a NativeAlarmSources field change. Proven by an
end-to-end flatten test: an instance native-alarm-source override that dangled
pre-fix now resolves against the re-imported template source.

Claude-Session: https://claude.ai/code/session_01MtdgwpEeCUn6cUA5f1LMPj
2026-07-09 16:59:37 -04:00
Joseph Doherty d833219920 fix(template-engine): DiffService covers native alarm sources — Deployments diff view no longer blind to native-alarm edits
Claude-Session: https://claude.ai/code/session_01MtdgwpEeCUn6cUA5f1LMPj
2026-07-09 16:46:45 -04:00
Joseph Doherty 0302694f9a fix(template-engine): native alarm sources participate in the revision hash — staleness detection restored (deliberate one-time stale flip for affected instances)
TemplateEngine-011: ResolvedNativeAlarmSource bindings were absent from the
flattened-config revision hash, so edits to a native alarm source binding
(source reference, connection, condition filter, locked state) went undetected
by staleness detection. Add a HashableNativeAlarmSource record and fold the
sorted, null-when-empty NativeAlarmSources collection into HashableConfiguration
in its alphabetical slot. WhenWritingNull keeps every native-source-free config
byte-identical to its pre-migration hash (pinned literal test proves it), so the
migration is surgical: only native-source-bearing instances flip stale once
after upgrade; redeploy clears it — deliberate.

Claude-Session: https://claude.ai/code/session_01MtdgwpEeCUn6cUA5f1LMPj
2026-07-09 16:46:45 -04:00
Joseph Doherty bf17f60a04 fix(transport): transport LockedInDerived on template attributes/alarms/scripts (additive DTO fields)
Adds a trailing optional `bool LockedInDerived = false` to TemplateAttributeDto,
TemplateAlarmDto and TemplateScriptDto (mirrors the ExecutionTimeoutSeconds
additive precedent — no schemaVersion bump). Populated at export
(ToBundleContent), consumed on import (FromBundleContent, BuildTemplate, and all
three SyncTemplate*Async changed-predicate + copy + new-entity paths, incl. audit
payloads). Old-form bundles that lack the property still deserialize to false.

Claude-Session: https://claude.ai/code/session_01MtdgwpEeCUn6cUA5f1LMPj
2026-07-09 16:46:33 -04:00
Joseph Doherty 874e32a93e fix(template-engine): collision detection sees all slots of a repeatedly-composed template
Claude-Session: https://claude.ai/code/session_01MtdgwpEeCUn6cUA5f1LMPj
2026-07-09 16:29:00 -04:00
Joseph Doherty 57f7f65772 fix(security): enforce template locks on native alarm source overrides at flatten and management-command level
Add IsLocked to ResolvedNativeAlarmSource; FlatteningService skips overrides on a
locked source (mirrors attribute/alarm lock rules); ManagementActor SetInstanceNativeAlarmSourceOverride
flattens the instance and rejects an override on a locked source (and rejects a
canonical name that does not resolve, preventing dangling overrides).

Claude-Session: https://claude.ai/code/session_01MtdgwpEeCUn6cUA5f1LMPj
2026-07-09 16:28:52 -04:00
Joseph Doherty ab6077708f fix(deployment): delete NotDeployed instances centrally without a site round-trip
A NotDeployed instance has no live Instance Actor at any site, so Delete is a pure
central-side record cleanup. Transport-imported instances always land NotDeployed
(often against an uncommissioned/unreachable site); a site round-trip made them
undeletable until the site came online. Short-circuit the site call for NotDeployed.

Claude-Session: https://claude.ai/code/session_01MtdgwpEeCUn6cUA5f1LMPj
2026-07-09 16:22:29 -04:00
Joseph Doherty 53ce64ba79 fix(transport): cap LineDiffer input size — oversized script diffs degrade to summary instead of O((N+M)^2) memory
Claude-Session: https://claude.ai/code/session_01MtdgwpEeCUn6cUA5f1LMPj
2026-07-09 16:13:05 -04:00
Joseph Doherty ef59c752ba perf(template-engine): cache script compile verdicts by code hash — unchanged scripts compile once per process
Claude-Session: https://claude.ai/code/session_01MtdgwpEeCUn6cUA5f1LMPj
2026-07-09 16:13:05 -04:00
Joseph Doherty f40f52950e fix(transport): export instance AreaName — Area-by-name reconciliation was half-shipped (exporter hardcoded null)
Claude-Session: https://claude.ai/code/session_01MtdgwpEeCUn6cUA5f1LMPj
2026-07-09 16:13:05 -04:00
Joseph Doherty 721a2e595f fix(deployment): audit-write fault can no longer flip a committed Success deployment to Failed
Route all post-terminal-status lifecycle audit writes (Deploy/Disable/Enable/Delete/DeleteOrphaned) through a swallow-and-warn TryLogAuditAsync helper so a failed best-effort audit write can neither reach the outer catch (flipping Success to Failed) nor propagate to the caller.

Claude-Session: https://claude.ai/code/session_01MtdgwpEeCUn6cUA5f1LMPj
2026-07-09 16:12:56 -04:00
Joseph Doherty 5c52b4c80b fix(transport): carry script cadence/timeout fields through bundle import (was silently reset to defaults)
Claude-Session: https://claude.ai/code/session_01MtdgwpEeCUn6cUA5f1LMPj
2026-07-09 16:09:50 -04:00
Joseph Doherty aae0a2db3e fix(transport): reject bundle imports that would persist an inheritance/composition cycle
Claude-Session: https://claude.ai/code/session_01MtdgwpEeCUn6cUA5f1LMPj
2026-07-09 15:49:55 -04:00
Joseph Doherty 854a6a8c0b perf(deployment): staleness/comparison paths no longer Roslyn-compile every script — deploy gate remains the authoritative compile
Claude-Session: https://claude.ai/code/session_01MtdgwpEeCUn6cUA5f1LMPj
2026-07-09 15:49:55 -04:00
Joseph Doherty 730bf19162 fix(template-engine): repeated composition no longer drops nested members — cycle guard keyed on recursion path, not global template id
Claude-Session: https://claude.ai/code/session_01MtdgwpEeCUn6cUA5f1LMPj
2026-07-09 15:49:55 -04:00
Joseph Doherty d19ab93d39 fix(security): deny Environment/GC and ADO.NET provider namespaces; close reflection-gateway list (GetTypes/EntryPoint/Declared*/DynamicInvoke)
Claude-Session: https://claude.ai/code/session_01MtdgwpEeCUn6cUA5f1LMPj
2026-07-09 15:49:55 -04:00
Joseph Doherty 6567216e14 perf(notification-outbox): bounded parallel delivery within a dispatch sweep (default 4) 2026-07-09 09:41:48 -04:00
Joseph Doherty acd8a6b8b5 chore(data-layer): low-severity cleanups — SQLite param chunking, KPI metric catalog, computed-column predicate guard 2026-07-09 09:32:04 -04:00
Joseph Doherty b58d810dbe fix(purge): short first tick on all three purge timers so daily-recycled nodes still purge 2026-07-09 09:26:42 -04:00
Joseph Doherty 34337a8714 perf(site-call-audit): OPTION (RECOMPILE) on the optional-parameter query page 2026-07-09 09:12:50 -04:00
Joseph Doherty 6f5d60070c perf(notification-outbox): single-query KPI aggregation; oldest-pending no longer materializes the live queue 2026-07-09 09:12:00 -04:00
Joseph Doherty 71189298a2 perf(kpi-history): time-sliced batched purge replaces the single daily mega-DELETE 2026-07-09 09:06:42 -04:00
Joseph Doherty bfd2cc7e6f fix(audit-log): reconciliation pulls fail over to site NodeB when NodeA is unreachable
SiteEntry gains an additive FallbackGrpcEndpoint; SiteEnumerator now picks NodeA
else NodeB as primary (a NodeB-only site is no longer skipped) and carries a
distinct NodeB as the fallback. Both GrpcPull{AuditEvents,SiteCalls} clients dial
the primary first and, on a transport fault (Unavailable/DeadlineExceeded/
Cancelled/HttpRequestException/SocketException/OperationCanceled), fail over to
NodeB once before collapsing to empty — a mapping/unexpected fault does NOT fail
over. Keeps the reconciliation loss-recovery net available during a NodeA outage.
2026-07-09 08:57:06 -04:00
Joseph Doherty 0385942c3f feat(audit-log): permanently-abandoned reconciliation rows leave a durable ReconciliationAbandoned audit record
When a pulled AuditEvent fails to insert on every retry up to the permanent-
abandon threshold, the reconciliation actor now writes ONE synthetic
ReconciliationAbandoned row (new AuditKind) alongside the Critical log line —
fresh EventId, Status=Failed, channel preserved from the lost row, Extra carrying
the abandoned EventId + source site + final error — via the same
ScadaBridgeAuditEventFactory the ingest path uses. Best-effort in its own
try/catch so it never blocks the cursor twice. The permanent loss is now
queryable in the Audit Log, not only in a rotating log file.
2026-07-09 08:50:44 -04:00
Joseph Doherty 20098c6108 fix(site-call-audit): composite keyset cursor eliminates the single-timestamp reconciliation pin; pinned state now a published event
Reconciliation cursor becomes composite (UpdatedAtUtc, TrackedOperationId);
IPullSiteCallsClient/GrpcPullSiteCallsClient forward the Task-15 after_id keyset
(additive param, null preserves the legacy inclusive >= contract). A burst
sharing one exact UpdatedAtUtc now drains via the id tiebreak instead of pinning
forever. A legacy site that ignores after_id is latched + published as
SiteCallReconciliationPinnedChanged on the EventStream (transition-only),
replacing the prior silent log line.
2026-07-09 08:44:07 -04:00
Joseph Doherty a608d7a79b feat(site-call-audit): additive (UpdatedAtUtc, TrackedOperationId) keyset in the reconciliation pull contract
Site side: additive proto after_id field, ReadChangedSinceAsync gains an optional
afterId cursor and deterministic (UpdatedAtUtc, TrackedOperationId) ordering — a
batch fully inside one UpdatedAtUtc instant no longer re-reads the same page
forever. Absent/empty afterId preserves the exact legacy inclusive >= contract, so
an older central is unaffected. Regenerated the checked-in gRPC C#.
2026-07-09 08:31:42 -04:00
Joseph Doherty 2c45c3238b feat(site-call-audit): relay Retry/Discard emit operator-identity audit rows
Additive RequestedBy on Retry/DiscardSiteCallRequest, plumbed from the Central UI.
On an Applied relay, SiteCallAuditActor emits one best-effort CachedResolve central
direct-write row (Submitted/Discarded) with the operator as Actor. Gated on an
injected ICentralAuditWriter (null in existing tests → no mirror read, unchanged);
the site remains the source of truth for the state change.
2026-07-09 08:21:41 -04:00
Joseph Doherty 50bb1ef8ab feat(notification-outbox): operator Retry/Discard emit audit rows with operator identity
Additive RequestedBy on Retry/DiscardNotificationRequest, plumbed from the Central
UI. RetryAsync emits a NotifyDeliver Submitted row (records who un-parked); Discard
stamps the operator on the Terminal row. Best-effort — audit failure never aborts
the action.
2026-07-09 08:12:47 -04:00
Joseph Doherty 1dd89d523b fix(config-db): EnableRetryOnFailure fleet-wide + execution-strategy wrap of the combined-telemetry dual-write
Installs SqlServerRetryingExecutionStrategy (5 retries, 30s cap) so transient SQL
faults retry transparently on every read-side path. Manual transactions aren't
auto-retried, so AuditLogIngestActor's idempotent dual-write is wrapped in an
explicit CreateExecutionStrategy().ExecuteAsync to become retriable. Verified
empirically that existing manual transactions (BundleImporter) are unaffected —
EF skips the strategy while a transaction is already active.
2026-07-09 08:03:30 -04:00
Joseph Doherty 6b06d1efcb fix(site-call-audit): same-rank freshness tiebreaker — retrying calls no longer freeze RetryCount/LastError at first write
Within an equal NON-terminal rank the newest UpdatedAtUtc wins, unfreezing the
Attempted-phase RetryCount/LastError/HttpStatus. Terminal ranks (>=3) are
excluded from the tiebreaker so terminal immutability is preserved (Delivered
never overwrites Parked); equal stamps stay idempotent, lower rank stays a no-op.
2026-07-09 07:56:39 -04:00
Joseph Doherty 1b53de1933 fix(kpi-history): backlogTotal trend records the real site-backlog aggregate instead of a hardwired zero 2026-07-09 07:51:55 -04:00
Joseph Doherty 5d0fcd4e66 perf(audit-log): bound GetExecutionTreeAsync edge scan to a root-anchored time window (partition elimination) 2026-07-09 07:10:35 -04:00
Joseph Doherty f1957606ff perf(site-call-audit): filtered IX_SiteCalls_NonTerminal index — KPI predicates seek the live queue instead of scanning 365 days 2026-07-09 07:06:50 -04:00
Joseph Doherty 7555e65746 fix(config-db): grant CREATE TABLE + scoped DELETE to scadabridge_audit_purger so a segregated maintenance principal can actually run the purge 2026-07-09 06:52:44 -04:00
Joseph Doherty 14c4df54f9 feat(audit-log): surface partition-purge failure as health event + counter 2026-07-09 06:49:28 -04:00
Joseph Doherty b59aa2d717 fix(audit-log): explicit long CommandTimeout on partition-switch and per-channel purge maintenance paths 2026-07-09 06:43:23 -04:00
Joseph Doherty 74aa872c07 feat(audit-log): daily site SQLite retention purge job (closes unbounded site DB growth)
SiteAuditRetentionService (site-only IHostedService) ticks PurgeExpiredAsync on
InitialDelay(5m)/PurgeInterval(24h, clamped >=1m), cutoff = UtcNow - RetentionDays.
Registered site-only in AddAuditLogHealthMetricsBridge; options bound from
AuditLog:SiteRetention. Per-tick failures swallowed. Added a test-only
PurgeIntervalOverride (mirrors SiteCallAuditOptions) so the ms-cadence tick test
can bypass the 1-min production clamp. Doc updated. (PLAN-04 Task 3, S1/U2)
2026-07-09 06:22:19 -04:00
Joseph Doherty b5271da5c8 feat(audit-log): SiteAuditRetentionOptions (7-day default, clamped 1-90)
RetentionDays clamped [1,90]; PurgeInterval clamped to >= 1 min (Timer spin
footgun); 5-min InitialDelay so a daily-recycled node still purges. (PLAN-04 Task 2)
2026-07-09 06:13:05 -04:00
Joseph Doherty 731ee69797 feat(audit-log): add site SQLite PurgeExpiredAsync honoring the ForwardState invariant
Retention purge deletes Forwarded/Reconciled rows older than the cutoff and
reclaims pages via incremental_vacuum; Pending rows are never purged on age
alone (would lose audit data that never reached central). Temp-table purge set
avoids the SQLite 999-param limit; runs under _writeLock. (PLAN-04 Task 1, S1/U2)
2026-07-09 06:11:54 -04:00
Joseph Doherty c9ac075e39 fix(site-runtime,dcl): low-severity cleanups — observed adapter dispose (S9), alarm-filter overwrite warning (S10), bounded-channel comment (C2), invariant-culture condition fallback (C6) 2026-07-09 02:09:06 -04:00
Joseph Doherty 29ee9176a7 feat(site-runtime): wire ConfigFetchRetryCount into the standby replicated-config fetch (UA2) 2026-07-09 01:26:58 -04:00
Joseph Doherty c457e8f464 feat(site-runtime): cert-trust reconcile-on-join — singleton pushes trusted certs to (re)joining site nodes (UA1) 2026-07-09 01:22:13 -04:00
Joseph Doherty 39976772d9 feat(site-runtime): CertStoreActor exports trusted certs (thumbprint+DER) for node reconciliation (UA1 groundwork) 2026-07-09 01:15:51 -04:00
Joseph Doherty 2efe1a4202 feat(site-runtime): persist native-alarm display metadata so rehydrated conditions render fully (UA4) 2026-07-09 01:09:13 -04:00
Joseph Doherty 326d945d85 perf(site-runtime): coalesce native-alarm mirror persistence into batched flushes (P4) 2026-07-09 01:05:20 -04:00
Joseph Doherty d2feb92bfd fix(site-runtime): site SQLite runs WAL with an explicit busy-timeout floor (S8) — concurrent writers stop serializing on the rollback journal 2026-07-09 01:00:12 -04:00
Joseph Doherty d124025cc7 perf(site-runtime): per-attribute child routing in InstanceActor — attribute changes reach only interested children (P2) 2026-07-09 00:55:26 -04:00
Joseph Doherty 1673268aee perf(dcl): tag->instances reverse index replaces per-update all-instances scan (P3) 2026-07-09 00:48:22 -04:00
Joseph Doherty f294b46263 fix(dcl): capture adapter into locals before background tasks — no off-thread reads of the mutable _adapter field (S7) 2026-07-09 00:44:09 -04:00
Joseph Doherty 957db62df1 fix(site-runtime): deployment whose Instance Actor dies during init reports Failed, exactly-once reply (S6)
Success now requires BOTH persistence commit AND an InstanceActorInitialized
readiness signal from the actor's PreStart — persistence can commit before the
actor's async init has run or failed, so persistence alone must not report
Success. An actor that dies during init never signals readiness; the Terminated
fallback fails that deployment and rolls back the optimistic state. The
persisted-row rollback is deferred until the store commits so it cannot race the
optimistic write.

Deviation from plan Task 15: the plan's swallow-only guard did not handle the
persistence-first ordering (empirically the store commits before the Terminated
signal, so Success was reported for a dead actor). Added the readiness handshake
to make the join deterministic.
2026-07-09 00:33:40 -04:00